Compare commits

..
Author SHA1 Message Date
Thomas Camlong 2ebfe659b2 revert(perf): drop the dockerode and docker-modem pnpm patches
Both patches came out of this work and both are removed at the maintainer's
request. The finding was real - ssh2's WASM heap costs +14.0 MiB RSS on require,
and @grpc/grpc-js another 14.6 MiB across 61 files, neither of which a
socket-based Docker install ever uses - but patching a dependency means every
future dockerode or docker-modem bump fails until someone regenerates the patch,
and that recurring cost is not worth roughly 25 MiB.

The finding stays on record in SERVER-MEMORY.md because the fix belongs upstream:
both packages require a module at file scope while using it in exactly one
already-guarded branch, so a PR to each would fix it for every consumer with no
patch to maintain here.

The two remaining patches, @types/node-unifi and trpc-to-openapi, predate this
work and are untouched. allowUnusedPatches was already set on dev and is left
alone. Lockfile regenerated: patchedDependencies now lists only those two.

Also sharpens the Cache Components entry. The SqliteError was simply an unmigrated
build database, which is fixable - but feeding the build a database would be the
wrong fix, because serverSetting is per-installation data on the user's volume and
a shell prerendered against the build machine's database would bake one install's
settings into every other install's first paint. The error was Next.js correctly
reporting a request-time read sitting in the shell path.
2026-08-14 11:16:55 +02:00
Thomas Camlong ca0bb36664 docs(perf): Cache Components tested and rejected, with the numbers
Enabled cacheComponents: true on Next 16.3.0-preview.9 and built. The flag works
and the usual migration friction is absent - zero files export dynamic,
revalidate or fetchCache.

Two blockers. The [locale] dynamic root segment needs generateStaticParams,
which is cheap: the guide is explicit that one value suffices, because
dynamicParams: true now serves the App Shell for unlisted params and upgrades in
the background. Not 53 routes x 38 locales.

Past that, 111 routes fail with the same error - SqliteError: no such table:
serverSetting. Prerendering runs at build time and the root layout
unconditionally awaits auth, two DB reads and a cookie read. For a self-hosted
app whose database is created at runtime on the user's volume there is nothing to
query during a build. Deferring those reads means moving the providers that wrap
children inside Suspense, which leaves a static shell of little more than
<html><body> - the exact failure the migration guide warns about.

And the prize is ~16-21 ms: measured TTFB is ~3 ms for a route that skips the
root layout against ~19-24 ms for one that goes through it. Meanwhile the board -
the main screen - is already dynamic(..., { ssr: false }) with a skeleton, so it
already paints an instant shell and Cache Components has no server render to
unblock there.

Recorded with both caveats: the ~20 ms is warm local SQLite and would be larger
on a network database, and the useful idea inside this - not blocking the first
byte on auth plus two DB queries - can be had by streaming those reads in the
existing layout, without the migration.

Experiment fully reverted; no code changes in this commit.
2026-08-14 11:11:50 +02:00
Thomas Camlong 839b4721fb perf(iframe): lazy-load the iframe widget
Also from PR #6582. An iframe is an entire browsing context, so one scrolled off
a tall board is the single most expensive thing a widget can load eagerly - and
this one had no loading attribute at all.

Records what else was mined from that PR and declined, notably its dynamic()
for BoardItemMenu: that component already early-returns null outside edit mode,
so there is no per-item overlay cost, and dynamic() fetches its chunk when the
component renders rather than when it returns something - so the chunk would
still load on every board render. It only pays off with the isEditMode guard
hoisted to the call site, which is a board-rendering change worth measuring
first rather than adopting on faith.
2026-08-14 11:04:08 +02:00
Thomas Camlong 40374803d4 perf(widgets): stop building collapsed Accordion panels
From reading PR #6582, which adds keepMounted={false} to the two coolify
accordions. The widget sweep had checked Tabs and never checked Accordion, so it
missed this on all five sites - an honest gap that PR closed for two of them.

Accordion behaves exactly like Tabs, verified in the installed version rather
than assumed: with three marker children per panel, the default renders 3 of 3
children of the *collapsed* panel and keepMounted={false} renders 0.

It matters more here than in Tabs, because collapsed is an accordion section's
normal state rather than the exception:

  coolify instance-card          three remote-data lists: servers, applications,
                                 services - all built while collapsed
  coolify single-instance-layout same three lists
  health-monitoring cluster      per-node resource tables   (beyond #6582)
  firewall                       interface rows             (beyond #6582)

Left mounted: the multiReleasesRepositories input, which holds selection state
and is edit-modal only rather than on the board render path.

In every fixed case the open/closed state lives in the parent, so unmounting a
collapsed panel discards nothing.

The spec is renamed tabs-keep-mounted -> keep-mounted since it now pins both
components, and three verdicts in the audit table are corrected: coolify was
recorded "clean (bounded)" and firewall triaged on other grounds, while both were
carrying this cost.
2026-08-14 11:02:13 +02:00
Thomas Camlong 2de3c86b31 fix(docs): exclude generated output from the docs typecheck
CI failed @homarr/docs:typecheck with a wall of TS1473 "An import declaration can
only be used at the top level of a module" pointing at
build/__server/assets/js/*.js - Docusaurus's own emitted bundles, not source.

The docs tsconfig includes **/*.js with allowJs enabled and excludes only
node_modules, so anything the build writes into the package is typechecked as
source. The turbo `typecheck` task dependsOn ["^topo"] and not on its own
package's `build`, so `docusaurus build` and `tsc --noEmit` run concurrently -
and tsc can observe the transient build/__server/ directory before the build
removes it again.

That makes it a latent race rather than a deterministic failure, which is why it
has not bitten before: it needs @homarr/docs to be in the affected set, and
merging dev brought apps/docs changes in. Locally it passes either way, because a
completed build leaves no __server directory behind.

Verified against a control rather than assumed: with build/__server/assets/js
containing a nested import declaration, the current tsconfig reports 0 errors and
the same tsconfig without the exclude reports the 2 TS1473 errors, matching CI.
2026-08-13 19:08:08 +02:00
ajnart be0043ab81 style: auto-fix lint and format 2026-08-13 16:41:32 +00:00
Thomas Camlong 89f081538c Merge origin/dev into feat/phase1-perf-impl
The branch was 12 commits behind dev and GitHub reported it CONFLICTING, which is
why the real CI workflow had not run since 2026-08-10 - a conflicting PR has no
mergeable ref, so pull_request-triggered workflows are skipped. Only the
lightweight checks (semantic title, CodeQL, CodeRabbit) were still reporting,
which made the PR look green while nothing was actually being tested.

Three conflicts, all of them this branch's barrel-diet work meeting dev's new
What's Up Docker integration. In every case the branch's structure is kept and
WUD is absorbed into it rather than the reverse:

  integrations/base/creator.ts   keep the lazy loader map; add wud as
                                () => import(...) rather than dev's static import,
                                which would have undone the lazy creator
  integrations/src/index.ts      keep the slim barrel; re-export WudStats and
                                WudContainerUpdate as `export type`, which is
                                erased at build and so adds no runtime module
  widgets/src/index.tsx          stays deleted - this was the barrel that put
                                every widget in the initial client graph

Registering WUD in the slimmed structures then needed three more entries that the
merge could not infer: base/kind-map.ts (type-only kind -> class map),
widgets/src/import.ts (type-only WidgetImports map, both the import and the
member). Each was found by typecheck rather than guessed at.

Verified: 37/37 packages typecheck, 1866/1881 unit tests pass. The 3 failures are
home-assistant and pi-hole, both testcontainers-based; both pass 11/11 in
isolation and fail only under 150-file parallel load with the container's payload
half-initialised.
2026-08-13 18:40:08 +02:00
Thomas Camlong fd351e4adf test(widgets): pin the Tabs behaviour that keepMounted={false} depends on
health-monitoring and media-missing pass keepMounted={false} because Mantine
otherwise builds the inactive panel's whole subtree. That came from reading
Mantine's docs, which is a weak basis for a prop that trades a remount cost for a
mount saving: if Mantine ever stopped rendering inactive panels, the prop would be
pure loss and nobody would notice.

Verified in the installed version instead. With three marker children per panel,
the default renders 3 of 3 children of the *inactive* panel into the DOM, and
keepMounted={false} renders 0. If that changes, this fails and the prop should
come back out.

Also records two things checked while re-reading the audit:

_inputs/ is excluded from the widget sweep on purpose, not by oversight - its
2,748 lines are option editors reached only from the edit modal and the
custom-widget form, and the one widget importing from it uses `import type`,
which is erased at build.

The "statically analyzable dynamic imports" rule has exactly one exception,
translation/src/mapping.ts building import(`./lang/${language}.json`) over 36
languages. It stands: the rule targets accidental over-inclusion, and every
language is genuinely reachable since the user can switch at runtime. Server-side
message loading is already memoised per locale, so it is not per-request either.
2026-08-13 18:28:29 +02:00
Thomas Camlong 43cc40eb66 docs(perf): optimizePackageImports is inert here - Turbopack makes it a no-op
Seven Mantine entry points were missing from optimizePackageImports while
core/hooks/icons-react were listed. Added all seven, rebuilt, and the client
bundle came out byte-for-byte identical: 482 files, 18050.7 KiB, same largest
chunk, same chunk hashes.

Identical output down to the hashes normally means a cached build, and
turbopackFileSystemCacheForBuild is enabled here, so it got checked rather than
written up as a null result. The Next.js docs are explicit: "This is not needed
when using Turbopack. Turbopack automatically analyzes and optimizes imports
without requiring this configuration." The build banner confirms Turbopack.

So the change is reverted and the existing three-entry list documented as dead
config, kept only as a webpack fallback. This also retires the sub-question of
whether @homarr/ui - the largest barrel at 111 importers - could be optimised
this way: it cannot, and it does not need to be.

Two notes for anyone revisiting: Next 16 with Turbopack prints no size columns in
its route table and emits no app-build-manifest.json, so per-route first-load JS
is unavailable from build output; these numbers are total bytes under
.next/static/chunks.
2026-08-13 15:36:12 +02:00
Thomas Camlong 320c51ac89 fix(http): stop destroying superseded agents, which could abort live requests
Follow-up to the agent caching change. When the trust material changes the caches
are cleared, and that path called agent.destroy() on the node https agents.

Node's documentation for agent.destroy() is "Destroy any sockets that are
currently in use by the agent" - so an admin uploading a certificate could have
aborted an HTTPS request that was already in flight, for instance a nextcloud
calendar sync. The agents run without keepAlive (verified: new https.Agent({})
reports keepAlive false on node 24), so destroying bought nothing anyway; each
socket closes on its own once its response completes.

undici's close() is kept, because it waits for enqueued requests to finish and
therefore cannot interrupt one in flight.

Also verified the correctness premise of sharing a single agent across every
integration, which the previous commit asserted but did not test: three origins
served by one agent each got their own pooled connection, four requests over one
connection apiece, with no response misrouted between them.
2026-08-13 15:36:12 +02:00
Thomas Camlong 09df2910c5 docs(perf): check the widgets against Vercel's React/Next performance rules
Sourced with the skills CLI - npx skills find "react performance", then
npx skills use affaan-m/ecc@react-performance, an adaptation of vercel-labs
react-best-practices: 70+ rules across 8 categories. Every mechanically
checkable rule was run against the widgets.

Seven already hold, and that is worth recording as much as a finding: no widget
registers a scroll listener, no numeric && can render a literal 0, the nested
array scans that exist are over single-digit collections, React.cache() is
applied, TanStack Query does the deduplication, content-visibility is on the six
lists that need it, and the two components-declared-in-render were already found
and fixed.

One rule this codebase deliberately contradicts: the skill prefers React 19
<Activity> over unmounting for tabs, and this pass did the opposite with
keepMounted={false}. The rule optimises for switching; a dashboard optimises for
mounting. Every board load mounts every widget while a user switches a widget's
inner tabs rarely, so building a hidden monitoring view on every page load buys
a faster switch nobody performs. Where state loss would show - custom-api's user
JSX, the edit modal's forms - the panels stay mounted.

Two candidates declined rather than guessed at, both recorded with reasoning:
the 20+ `data = []` defaults only destabilise while data is undefined, since
react-query returns a stable reference once loaded; and the 7 unlisted Mantine
entry points need two production builds and a first-load-JS diff before anyone
claims they help. @homarr/ui is not eligible at all despite 111 importers,
because Next's barrel optimiser only rewrites pure re-export files.

Also records the process error behind the earlier "no skills CLI" claim: the
entries in ~/.claude/skills are symlinks, which find -type f skips, so the
directory looked empty right after cat had succeeded on a file inside it.
2026-08-11 16:40:27 +02:00
Thomas Camlong 083a926719 docs(perf): give every one of the 50 widgets a verdict
Completes the sweep: 11 fixed, 8 triaged, 10 read and found clean, 21 verified
bounded. Each verdict records the evidence behind it, because "I read this" and
"its list cannot get long" are different strengths of claim and should not look
alike in a table.

The four widgets the ranking flagged as most likely to hide something all came
back clean, and three were already bounded before I looked - traefik slices to
what its width shows, umami to its own limit option, beszel-alerts to
options.maxHistoryItems. system-resources looked worst of all with 14 unmemoised
maps, until it turned out none of the six chart components it feeds is memoised
either, so a stable array identity would preserve nothing.
2026-08-11 16:33:33 +02:00
Thomas Camlong ec79a15023 perf(widgets): stop building hidden tab panels, and drop a third-party image fallback
Mantine's Tabs keeps inactive panels mounted by default. On Mantine 9 with
React 19 that uses <Activity>, which preserves state and unmounts effects but
still builds the entire subtree - so a two-tab widget pays for both while
showing one.

  health-monitoring  both the system and the cluster monitoring view were built
  media-missing      both tabs built their full card list, each card with a
                     poster and a blurred backdrop

Two other Tabs users are deliberately left mounted: custom-api panels can hold
user component state, and the widget edit modal would reset half-filled forms.

media-missing also rendered every card into a scrolling tile with no lazy
loading, the same shape as media-releases. Its card height is fixed per density,
so contain-intrinsic-size is exact rather than a guess.

calendar fell back to https://placehold.co/400x400 when a poster failed to
load, so a self-hosted dashboard made a third-party request - once per failing
event - to draw a grey square, and showed nothing at all on a host without
internet access. Same grey square, inline.

The rss feed handler logged the whole serialised feed entry through a raw
console.debug, which no log level could suppress, and reported a match count
that was really the regex capture-group count.

Adds scripts/benchmarks/widget-inventory.mjs and docs/perf/WIDGET-AUDIT.md, a
verdict for all 50 widgets.
2026-08-11 16:30:22 +02:00
Thomas Camlong cf240addea perf(image-proxy): stream proxied images instead of buffering them
forwardImageAsync awaited response.arrayBuffer(), so the whole image sat in
server memory before anything was sent to the client. A media board asks for
dozens of posters and backdrops at once, so that is dozens of full images held
simultaneously - and a client that gives up midway still cost a complete
server-side download.

Forwarding response.body instead lets backpressure reach upstream. Measured
with 30 concurrent 2 MiB images and deliberately slow clients, one fresh
process per arm because the allocator does not return freed spans:

  buffered   +116.0 MiB, +117.8 MiB peak RSS
  streamed    +74.2 MiB,  +78.8 MiB peak RSS

-36% peak RSS on that load, and it scales with concurrency times image size.
Content-Length is now forwarded too, so clients still get a progress bar.
2026-08-11 16:09:37 +02:00
Thomas Camlong 3ba2800db9 perf(http): reuse the certificate agent instead of building one per request
fetchWithTrustedCertificatesAsync has 125 call sites and none of them pass a
dispatcher, so every HTTP request Homarr makes - every integration poll, every
proxied image - built a brand new EnvHttpProxyAgent. An agent owns its
connection pool and its TLS session cache, so a fresh one per request means a
full TLS handshake every time and no keep-alive reuse at all.

Measured against a local TLS server, 60 sequential requests:

  fresh agent per request   5.11 ms/req   60 TLS handshakes
  one shared agent          1.69 ms/req    2 TLS handshakes

3.0x faster and 30x fewer handshakes on loopback; over a real network the gap
is larger, since that is where handshake round-trips actually cost something.

Agent construction itself is cheap (0.004 ms, 1.7 KiB) because undici builds
the TLS store lazily at connect time - so the cost was never construction, it
was the handshake that a fresh pool forces.

The trust material is still read on every call and the cached agents are keyed
on a hash of it, so adding or removing a certificate - including by dropping a
file into the mounted folder rather than going through the UI - takes effect on
the very next request, exactly as before. Superseded agents are closed so their
idle sockets are released.
2026-08-11 16:09:37 +02:00
Thomas Camlong 9b54edd2db docs(perf): correct the "16 MiB tRPC subtree" lever - it was already removed
This branch has repeatedly named a 16.10 MiB subtree of 460 files, reached by every API
route, as the one remaining server-side lever with real headroom, attributing it to the tRPC
router importing every integration eagerly. Both halves are wrong.

The integration creator was already fully lazy: every kind in creator.ts is
`() => import(...)`, so no integration class loads until one is constructed. The earlier
barrel diet had done this properly - the claim was never checked against the code.

The subtree was an artefact of route preloading, measured before preloadEntriesOnStart:
false shipped, when Next loaded all 74 routes' module graphs at startup. Re-measured on the
current build, boot is 706 files and 10.56 MiB:

     3.73 MiB    28 files  .next/server/chunks (the app's own boot graph)
     2.16 MiB    14 files  better-sqlite3 (2.1 MiB of it the native binary)
     1.75 MiB   359 files  next/dist
     1.58 MiB    25 files  next/dist/compiled
     0.45 MiB    77 files  mysql2        <- known, structurally blocked
     0.07 MiB    13 files  pg            <- same

There is no single large server lever left. What remains is diffuse: a full multi-tab session
loads 15.44 MiB of app chunks across 241 files, and both the ~41 MiB of retained source and
the ~38 MiB of compiler zones scale with that. Reducing it means shipping less application
JavaScript per route, which is a broad refactor with no single target.
2026-08-11 10:22:31 +02:00
Thomas Camlong 37dead1690 docs(perf): verify the off-screen changes, and record two instruments that lied
The off-screen work was committed unmeasured; this closes that debt. Measured on two
containers restored from identical data:

  image requests   138, 135  ->  22, 22   (-84%)
  image bytes        12.27 MiB -> 0.03 MiB
  layout objects  11611, 7836  ->  3694, 3705  (-53%)

A 99.8% drop in image bytes is not believable, so it was checked instead of reported. Two
checks were bad instruments before one worked:

  - Counting decoded images via getBoundingClientRect() reported 80/80 loaded in BOTH arms.
    Reading geometry forces layout, which forces content-visibility subtrees to render, which
    triggers the very image loads being tested for. The instrument caused the result.
  - Screenshots came back byte-identical in both arms, because this environment shows an
    onboarding overlay and media-requests renders blank with unreachable integrations. Good
    evidence of no visual difference; useless as evidence that images render.
  - Counting img.complete && naturalWidth > 0 reads no geometry and gave the answer: 80/80
    decoded in both arms, 138 responses before against 25 after.

So nothing is visually broken. The 80 imgs are integration icons sharing URLs, deduplicated to
~25 unique fetches in both arms; the ~113 avoided requests are other widgets' off-screen
artwork, chiefly media-requests' full-size per-card backdrops, which is also where the 12 MiB
came from.

Stated limit: media-requests renders blank in this environment, so its visible-card behaviour
after the change is unverified. Media-releases is verified at 80/80 decoded.
2026-08-11 10:20:33 +02:00
Thomas Camlong d2e86895f4 docs(perf): rewrite the client doc with the measured overlay results
Replaces the earlier version with what was actually measured, and closes out all 54
findings from the widget audit rather than leaving categories open:

  overlay-per-row (12)      5 fixed; 7 are functional tooltips with real labels
  index-key (11)            all safe - static text splits, or stateless rows from
                            user-configured JSONPath data with no stable id
  nested-component (2)      both fixed
  inline-style-per-row (29) left alone; a fresh object per row only defeats memoisation
                            if the child is memoised, and these are not

Measured, deterministic per-widget counts:

  mediaReleases      340 comp / 8387 fib  ->  170 / 8057
  calendar           171 comp / 1195 fib  ->   75 /  907
  beszelSystemTable   36 comp / 1286 fib  ->   18 / 1230

Records two metrics that turned out unusable - Chrome's board-level JSEventListeners read
7905 to 17114 across six runs of identical code, and its Nodes counter includes detached
nodes - and one claim withdrawn: the media-releases tooltip fix measured zero, the win for
that widget came entirely from the shared OverflowBadge fix.
2026-08-11 10:11:08 +02:00
Thomas Camlong afa68687eb perf(widget-inputs): stabilise the drag handle component type
Last of the two nested-component findings from the widget audit. `Handle` was declared in the
render body and passed to a child as `handle={Handle}` - a component *type*. A type that
changes identity makes React unmount and remount the subtree instead of updating it, and this
component re-renders on every frame of a drag as dnd-kit updates `transform`, so the handle was
being torn down and rebuilt continuously while being dragged.

Wrapped in useMemo keyed on the two things it closes over.

Measured result of the whole overlay/remount sweep, per widget, deterministic component and
fiber counts on identical restored data:

  widget                    before              after
  mediaReleases      340 comp / 8387 fib   170 / 8057   (-50% components)
  calendar           171 comp / 1195 fib    75 /  907   (-56% components, -24% fibers)
  beszelSystemTable   36 comp / 1286 fib    18 / 1230   (-50% components)
  downloads / mediaRequests / app        unchanged - their overlays are functional

Two corrections to earlier claims in this branch:

  - The media-releases win came from the shared OverflowBadge fix, NOT from the tooltip change.
    Measured on the build with only the tooltip fix, mediaReleases was still at 340; it dropped
    to 170 once OverflowBadge stopped mounting a Popover when nothing overflows. The tooltip fix
    remains zero on this board because every release has a description.
  - The board-level "event listeners" metric is unusable. Across six runs of identical code it
    read 7905, 8000, 8166, 8179, 10360 and 17114. Chrome's Nodes counter has the same problem
    (23072 on one run against ~8850 on the rest) because it counts detached nodes awaiting
    collection, while "elements under body" stayed at 6691-6726 throughout. Only the per-widget
    component and fiber counts are stable enough to compare, so those are what is reported.
2026-08-11 10:09:59 +02:00
Thomas Camlong 2534103627 perf(ui): stop mounting Popover machinery for overlays that cannot open
Two more instances of the same Mantine behaviour, found by searching for overlay components
that take a `disabled` prop rather than by measuring another board.

OverflowBadge (packages/ui) mounted a Popover and a Popover.Dropdown unconditionally - both
when `disablePopover` was set and, more importantly, when nothing overflowed at all, which is
the common case. It is shared UI, used by media-releases, releases, the docker table and the
widget inputs, so the cost repeated per row wherever a list renders it. The rendered markup is
unchanged: a Popover.Target only forwards a ref, so returning the badges directly is
visually identical.

beszel disk-usage mounted a HoverCard per table row with `disabled={filesystems.length === 0}`,
and already fell back to a plain progress bar inside its own target for that case - so for a
system reporting no extra filesystems the entire overlay existed to render nothing. Hoisted
that branch out; the now-unreachable ternary and the dead `disabled` prop went with it.

Measured result of this class of fix, on the calendar which had the worst ratio:

  interactive components  171 -> 75   (-56%)
  fibers                 1195 -> 907  (-24%)

Honest note on the media-releases tooltip fix committed earlier: it measured **zero** on the
test board. Every release there has a description and the option is on, so all 80 tooltips are
legitimately enabled and none are skipped. It only helps boards where descriptions are absent
or the option is off. The mechanism is right; this board simply is not the case it helps.

Also confirmed NOT worth changing: downloads mounts 30 Tooltips but all three per row are
functional (integration name, hover detail, progress), so there is nothing disabled to skip.
Replacing them with native title attributes would change the UX, not just the cost.
2026-08-11 10:04:12 +02:00
Thomas Camlong 365b2ed13e perf(calendar): stop mounting a HoverCard per day cell that can never open
Third instance of the same pattern, and the worst ratio yet. Mantine mounts a HoverCard's
Popover machinery - Popover, PopoverTarget and PopoverDropdown - even when it is disabled.
A month grid is around 42 cells and `disabled` is `isEditMode || eventsForDate.length === 0`,
so on a typical month most days mounted three components for a card that can never open, and
in edit mode every single day did.

Measured on a real board: 42 Popovers, 42 PopoverDropdowns and 42 PopoverTargets - 126 of
this widget's 171 listener-attaching components - in a widget with only 251 DOM nodes. That
is a far worse components-per-node ratio than media-releases, which had 340 across 2,951
nodes.

Returning the day cell directly when disabled is behaviour-identical: a disabled HoverCard
never opens, so nothing observable is lost.

This is now the third place this exact Mantine behaviour has cost real work - after the
shared board context menu and the media-release tooltips. The general rule: Mantine overlay
components mount their full machinery regardless of `disabled`, so `disabled` is a
visual-suppression prop, not a cost-avoidance one. Deciding before rendering is what avoids
the cost.
2026-08-11 09:58:13 +02:00
Thomas Camlong 7e961d3da4 perf(beszel-system-table): hoist PercentCell out of the render body
A static audit of every widget (scripts/benchmarks/audit-widget-render.mjs, added here)
found this: PercentCell was declared inside the widget body, which makes it a new component
type on every render. React cannot match a new type to the previous one, so it unmounted and
remounted every cell's subtree - Group, Text and Progress - instead of updating it. The
widget re-renders about once a second from its live Beszel subscription, and the cell is used
by three columns, so that was three subtrees per system torn down and rebuilt every second.

Hoisted to module scope and wrapped in memo. `size` is now passed as a prop; it was the only
thing the nested version closed over.

The audit covers all widget .tsx files and ranks findings by whether they sit inside a
.map(), since those multiply by row count:

    29  inline-style-per-row
    12  overlay-per-row
    11  index-key
     2  nested-component

Deliberately not fixing the other 52 yet. The 29 inline styles are the least likely to
matter - a fresh object per row only defeats memoisation if the child is memoised, and mostly
these are not - so changing 29 call sites would be churn without a measurement to justify it.
The 11 index keys are only harmful where a list reorders, which needs checking per widget. The
remaining overlay-per-row sites are the same pattern already fixed twice (context menus, then
media-release tooltips) and are the most likely to be worth it: calendar mounts 42 Popovers
and downloads 30 Tooltips, both confirmed at runtime.

Not measured: the mechanism is certain but the size depends on how often the widget renders
and how many systems are in the table.
2026-08-11 09:52:37 +02:00
Thomas Camlong f5b7818c53 perf(media-releases): stop mounting a Popover per card for tooltips that never show
Attributing listener-attaching components per widget found the same anti-pattern the
shared board context menu already fixed once, still present in list widgets:

  interact   nodes  fibers  kind / which components
       340    2951    8387  mediaReleases
                            UnstyledButton x90  TooltipFloating x80  Popover x80
                            PopoverDropdown x80  PopoverTarget x10
       171     251    1195  calendar
                            UnstyledButton x45  Popover x42  PopoverDropdown x42
                            PopoverTarget x42
        46     264     898  downloads
                            Tooltip x30

Mantine mounts a TooltipFloating's Popover machinery even when the tooltip is disabled, so
wrapping every card unconditionally produced 80 TooltipFloating, 80 Popover and 80
PopoverDropdown components - 240 of this widget's 340 listener-attaching components - for
tooltips that only show on hover, on cards that are 91% off-screen.

Deciding before rendering means a card with no description, or with the option off, mounts
none of it. Behaviour is unchanged for cards that do have a description.

Also extends measure-widget-cost.mts to count these components per widget, which is what
found it. The board's 8,165 event listeners were never attributable before; now they are.

Not yet measured: the listener reduction depends on how many releases carry a description,
which is integration data. The mechanism is certain (the components are simply not created)
but the size of the win on a given board is not, and calendar's 42 Popovers and downloads'
30 Tooltips are untouched. Tracked in docs/perf/NEXT-STEPS.md.
2026-08-11 09:42:59 +02:00
Thomas Camlong fee2e6a6af docs(perf): track which off-screen applications are measured and which are not
Four commits now apply the off-screen row treatment; only the first was A/B'd. Records
which is which, and warns how to read a result for the three widgets that do not overdraw
on the test backup - a measurement there will correctly show no change, because the benefit
only exists once an integration returns more rows than the tile can show.

Also records two things that look like gaps and are not:

  - content-visibility was applied to five of the 26 ScrollArea widgets on purpose. It
    costs containment work and can be wrong on self-measuring content such as charts and
    the notebook, so the rule was: apply where list length is decided by remote data.
  - React cache() is server-only. It already wraps getBoardByNameAsync, getHomeBoardAsync,
    getRscServerSettingsAsync and getQueryClient; it does nothing for client widget
    rendering, where react-query does the caching and its gcTime was already verified at
    5 minutes rather than the 24 hours the original fix catalogue assumed.
2026-08-11 09:38:59 +02:00
Thomas Camlong 040f105946 perf(widgets): share the off-screen row primitive, and apply it to the data-driven lists
Extracts the media-releases fix into packages/widgets/src/common/offscreen-rows.ts and
applies it to the remaining widgets whose list length is decided by remote data rather
than by the tile: rssFeed, notifications and the calendar event list. Images in those
lists also get loading="lazy" - the rss cards carry two each, a backdrop and a poster.

The measured basis, from media-releases where the pattern was found:

  image requests   135 -> 55    (-59%)
  layout objects  8346 -> 5117  (-41%)

Scope was chosen deliberately rather than applied everywhere. 26 widgets use ScrollArea,
but on the measured board only two rendered more than their tile could show, and
content-visibility is not free: it costs a little containment work, and on content that
measures itself - charts, editors - skipping layout can be actively wrong. The widgets
changed here are the ones where overdraw is a function of how much data an integration
returns, so a user with a busier instance hits it even though this board does not. Chart
and fixed-size widgets are deliberately left alone.

Honest scope note: unlike the media-releases change, these three are NOT measured. On the
test board rssFeed renders 7 nodes and the calendar's 7 rows fit their tile, so there is
nothing to improve here to observe - the benefit only appears with more data than this
backup has. They are typechecked and lint-clean. docs/perf/NEXT-STEPS.md tracks them
alongside the media-requests change as pending verification.

A duplicate `style` prop on the notifications Card was caught by typecheck and merged
rather than left to override the existing colour and text-decoration styles.
2026-08-11 09:37:48 +02:00
Thomas Camlong 8f94ea47e3 docs(perf): add a handoff with the unverified item, the remaining lever, and the clean-PR set
Records where this stopped so it can be picked up cleanly:

  1. c526dd2b7 (media-requests content-visibility) is typechecked but NOT measured - the
     only performance claim on this branch not A/B'd before commit. Verify with
     measure-board-paint.mts, three repeats; revert if image requests do not fall.
  2. The one lever with headroom left is the tRPC router's eager integration imports - a
     16.10 MiB subtree of 460 files reached by every API route, which pays twice because
     both the ~41 MiB of retained source and the ~38 MiB of V8 compiler zones scale with
     how much JavaScript the server compiles. Every allocator and V8-flag avenue is closed.
  3. Deferred with prizes attached: ASCII-escaping bundles (9.45 MiB, needs its own PR),
     lazy mysql2/pg (21.4 MiB, structurally blocked), beszelSystemStats chart SVG.
  4. The proposed set for the clean PR, six changes producing -26% peak anon and -52% at
     idle for +18% CPU - with the gRPC patch flagged as hygiene rather than a measured win,
     and the +30% board-load figure flagged as not established.
2026-08-10 19:12:28 +02:00
Thomas Camlong c526dd2b77 perf(media-requests): apply the off-screen fix to the request list — NOT YET MEASURED
Second instance of the pattern fixed in mediaReleases. The overdraw tool, corrected to
find row containers by child count rather than by uniform height, found only one other
widget genuinely rendering more than its tile can show:

  nodes  rows  vis  offscr  imgs   tile  content  kind
   2951    80    7    2693    80    539     7206  mediaReleases            (already fixed)
    393    20    6     275    60    405     1512  mediaRequests-requestList

1,512 px of cards in a 405 px tile, and 20 cards carrying 60 images between them - each
card renders a full-tile backdrop plus a poster. Same treatment: content-visibility with
contain-intrinsic-size on the card, loading="lazy" on both images.

IMPORTANT: this is typechecked but **not measured**. Every other performance claim in this
branch was verified by an A/B before being committed; this one was not, because the session
ended first. It is committed so the work is not lost, and it is the first thing to verify
next - the same measurement used for mediaReleases applies:

  STRESS_BASE_URL=http://127.0.0.1:PORT node --experimental-strip-types \
    scripts/benchmarks/measure-board-paint.mts

Expect image requests to fall; do not expect JS heap to move, since decoded bitmaps live
outside JSHeapUsedSize.

Every other widget on the board fits its tile and needs nothing. The tool's "visible"
column misreads chart widgets (beszelSystemStats reports 28 rows / 0 visible because its
container is an SVG whose groups have no layout height) - content height versus tile height
is the trustworthy signal, and by that measure only these two overdraw.
2026-08-10 19:11:38 +02:00
Thomas Camlong c73b139dbd docs(perf): split the findings into per-subject markdown files
PHASE2-FINDINGS.md had grown past 1,200 lines mixing server memory, client rendering,
methodology and a long list of rejected experiments, which made it unusable as a
reference. Split under docs/perf/:

  README.md              headline results and the change-by-change ratio table
  SERVER-MEMORY.md       what the container's RAM is made of, layer by layer
  CLIENT-PERFORMANCE.md  dashboard cost in Chrome: DOM, fibers, images, paint
  REJECTED.md            ideas implemented, measured and abandoned, with numbers
  METHODOLOGY.md         the tools, and the measurement traps that gave wrong answers

PHASE2-FINDINGS.md is now an index so existing links keep working.

METHODOLOGY.md is the part worth keeping: every wrong number in this investigation came
from an instrument rather than the application, and each trap is written down - snapshots
cannot explain peaks, V8 truncates snapshot strings to ~1 KiB, snapshot objects are not
live handles, measuring without a forced GC measures V8's growth, taking a CDP snapshot
added 134 MiB to the process being measured, per-component React attribution silently
returns nothing on a production build, a fixed port can measure a leftover container, and
a server-side change that moves a client-side metric means the workload changed.
2026-08-10 19:08:20 +02:00
Thomas Camlong 7550766f67 perf(media-releases): stop rendering 73 off-screen cards' images and layout
Per-widget attribution of the dashboard found one widget responsible for 44% of the
board's DOM: mediaReleases at 2,951 nodes and 8,387 fibers, including 1,935 Mantine
Box components, against a board total of 9,293 nodes.

The cause is that it maps over every release the integrations return, with no cap.
Measured on a real board: the tile is 539 px tall and its content is 7,206 px, so of
80 rendered cards only 7 are visible. The other 73 cost DOM, fibers, and - most
expensively - a poster and a CSS backdrop image each, all fetched and decoded for
pixels nobody can see. Decoded bitmaps are far larger in memory than their files.

Fixed with content-visibility: auto plus contain-intrinsic-size on each card, so the
browser skips layout, paint, raster and background-image fetch for off-screen cards,
and loading="lazy" on the poster. Nothing leaves the DOM, so scrolling and in-page
search are unchanged, and the intrinsic size keeps the scrollbar from jumping.

Measured over three repeats per side on identical restored data:

  image requests    135, 135, 135  ->  55, 55, 56   (-59%, deterministic)
  image bytes            12.27 MiB  ->  8.41 MiB    (-31%)
  layout objects  8346, 7815, 7815  ->  5117, 4696, 4636  (-41%)

Two things this does NOT show, stated because a single run suggested otherwise:

  - JS heap did not clearly improve. Across four runs it was 46-88 MiB before against
    74-86 MiB after; the before spread is 2x, so no claim either way. This is expected:
    decoded image memory lives outside JSHeapUsedSize, so the actual saving from 81
    fewer image fetches is invisible to that metric.
  - Task duration and time-to-settled overlap between arms and support no conclusion.

Adds the tools this needed: measure-widget-cost.mts (per-widget DOM nodes, fibers and
heaviest components), measure-board-paint.mts (image requests, bytes, style/layout/task
time, layout objects), measure-widget-overdraw.mjs (rendered versus visible rows).
2026-08-10 19:05:17 +02:00
Thomas Camlong 5fb8e9223a docs(findings): the 38 MiB of compiler zones is V8's optimiser, and not reachable by flags
Closes the largest unexplained item in the peak breakdown, then closes it off as a lever.

Attribution: peak_malloced_memory goes 4.5 -> 42.1 MiB the moment a board's modules
compile, while current malloced_memory stays at 1.0 - memory V8 takes to compile, frees
back to malloc, and musl never returns to the kernel. Confirmed by re-reading the
--jitless runs, where the spike never occurs at all (4.5 MiB at every stage).

--jitless is unusable because it under-renders the page, but V8 can cap the optimisation
tier instead of removing codegen. --max-opt is rejected by NODE_OPTIONS and accepted on
the command line, so it was wired behind HOMARR_MAX_OPT and A/B'd on one image:

  peak anon      276.0 -> 282.8 MiB  (+2%)
  boot idle       64.0 ->  72.3 MiB  (+13%)
  board loaded   186.1 -> 179.2 MiB  (-4%)
  4 tabs         270.8 -> 238.5 MiB  (-12%)
  after soak     199.0 -> 214.3 MiB  (+8%)
  CPU             28.1 ->  27.8 s    (unchanged)

Null result. Render integrity was verified explicitly this time - listeners 7,784 in all
four runs, DOM nodes within 40 - so unlike jitless the page rendered fully and the
comparison is valid.

The mechanism explains it: measured under --max-opt=2, peak_malloced_memory at board load
is 42.1 MiB, identical to Turbofan. Maglev's zones are just as large. Capping the tier does
not reduce zone memory; only not compiling does, and that breaks rendering.

So this 38 MiB is not reachable through V8 configuration. It scales with how much
JavaScript V8 must compile, which returns it to the single remaining server-side lever:
ship less server code. Reverted.
2026-08-10 18:43:31 +02:00
Thomas Camlong dc46057544 docs(findings): definitive server-RAM A/B, and jemalloc rejected with evidence
Two things this settles.

First, the end-to-end number. Every Phase 2 result so far was an incremental A/B against
the build immediately before it, which cannot be summed. Rebuilt the merge-base
(06724cfeb) from a clean worktree and ran 2+2 alternated against HEAD, reporting cgroup
anonymous memory - the figure that counts against a container limit and cannot be
reclaimed:

  stage              before     after
  peak anon        376.6 ->   279.3 MiB  (-26%)
  boot idle        130.5 ->    62.4 MiB  (-52%)
  after sign-in    247.4 ->   142.9 MiB  (-42%)
  dashboard        293.4 ->   170.4 MiB  (-42%)
  4 tabs           366.5 ->   258.2 MiB  (-30%)
  after soak       323.7 ->   200.8 MiB  (-38%)
  peak node RSS    430.7 ->   336.2 MiB  (-22%)

No before/after range overlaps at any stage. Costs: CPU 23.5 -> 27.7s (+18%, solid,
non-overlapping). Median board load reads +30% but is NOT established - individual loads
were [506,383,1233] and [848,431,574] before against [780,808,515] and [621,579,1004]
after, heavily overlapping with no cold-then-warm pattern, so six samples per side cannot
separate a real regression from the variance.

compare-stress-runs.mjs now leads with anonMiB, which the harness recorded all along but
the report was ignoring in favour of memory.current - a figure that includes page cache
and swings 60+ MiB between runs of a byte-identical image.

Second, jemalloc. The largest attributed-but-unfixed item is ~38 MiB of V8 zone memory
that musl frees to its own pools and never returns to the kernel, so an allocator that
decays spans back to the OS looked like the answer. Tested on the byte-identical image
with only HOMARR_MALLOC differing, asserting the "Using jemalloc" log line:

  boot idle    62.4 ->  80.3 MiB  (+29%)
  after soak  215.5 -> 279.6 MiB  (+30%)
  4 tabs      259.9 -> 319.6 MiB  (+23%)
  CPU          28.0 ->  24.5 s    (-13%)

The hypothesis was backwards. jemalloc keeps more freed memory on purpose - thread caches
plus time-based dirty-page decay - and spends it on speed. 13% cheaper CPU, 18-30% more
resident memory. Reverted, and it closes off "swap the allocator" as a RAM direction:
musl is already the frugal choice here.
2026-08-10 18:25:00 +02:00
Thomas Camlong f5a9933dd8 docs(findings): the Beszel coalescing fix measured zero, and was reverted
Second correction in the same area. Having found that three Beszel widgets account for
82% of idle DOM churn, the obvious fix was to batch their subscription updates instead
of calling setState per record. Implemented, built, measured:

  DOM mutations / 45s   733 -> 748
  React commits / 45s    55 ->  55

Identical. Beszel emits roughly one record per second and the flush interval was also
one second, so each flush batched exactly one record. Coalescing at the arrival rate
does nothing - arithmetic that should have come before the implementation.

That also exposes a unit error in the previous commit's numbers: DOM mutations are not
re-renders. The real rates are 1.22 React commits/sec for the whole board and 13.3 DOM
mutations per commit, because a chart redraw touches many nodes. The stats widget's
"10.4 mutations/sec" is ~0.8 renders/sec - three live-monitoring widgets each refreshing
about once a second, which is what they exist to do.

So 73 commits/min is not pathological, it is the feature working, and there is no defect
to fix here. Reverted rather than shipped: keeping speculative complexity that measured
zero - an extra timer plus unmount and system-change teardown - is the trade this work
rejects everywhere else.

What survives is narrower: idle churn is widget-driven rather than app-shell driven, and
legitimate. The 650 MB browser heap therefore comes from interaction and multi-tab use,
not idle churn, and remains unexplained.
2026-08-10 17:52:23 +02:00
Thomas Camlong 9f1a15b9bf docs(findings): the App Router churn claim was wrong; idle churn is Beszel
Corrects the largest wrong claim in this work. Earlier rounds reported that ~80% of
idle React commits entered at Next's internal HeadManagerContext and that "the App
Router root re-renders about once a second", and named it the highest-value next step.

Both are false. Measuring the same idle churn against boards of different sizes:

  board      items   idle commits/min
  hey            0                0.0
  footbol        3                0.0
  default       28               76.0

A 3-item board commits zero times per minute. If the router re-rendered on a timer,
every board would churn regardless of its contents.

The original attribution was a tooling artefact. measure-rerenders.mts decides a fiber
re-rendered from fiber.actualDuration, which only a React *profiling* build populates.
On a production build every fiber fails that test and both attribution tables come back
empty - and an empty table reads as "nothing re-rendered" rather than "this build cannot
tell you". The harness now says which case it is, and waits for the board container to be
attached rather than visible so a zero-item board can be measured at all - that control
is what disproved the claim.

Adds measure-widget-churn.mts, which attributes churn on a production build via DOM
mutations keyed to the nearest [data-type="item"] ancestor. On the 28-item board, idle
for 45s:

  470  beszelSystemStats      64%
   77  beszelSystemTable      11%
   52  beszelSystemGrid        7%
   22  dockerContainers        3%
   93  (outside any widget)   13%

Three Beszel widgets are 82% of all idle churn, one of them 64% on its own at 10.4
mutations per second.

Also logs what Next.js itself costs, since it was assumed to be the problem: its loaded
source is a flat 3.47 MiB (364 framework files + 32 vendored) and does not grow at all
between boot and a fully loaded multi-tab session, while Homarr's own chunks grow from
3.73 to 15.44 MiB. Next is 16.1 MiB on disk; the largest things it vendors are its
next-server runtime (1.60), @edge-runtime (0.85), edge-runtime (0.51) and compression
(0.30). The framework is not where this application's memory goes.
2026-08-10 17:45:12 +02:00
Thomas Camlong ab5b8dad43 docs(findings): attribute the native bucket at peak, and record two disproven theories
Closes the last unexplained part of the peak footprint. The 118.8 MiB 'other
anonymous' smaps bucket is now mostly accounted for: V8 large-object space 34.5,
external buffers 24.6, and ~38 MiB of V8 parser/compiler zone memory that was freed
back to malloc but never returned to the kernel by musl - peak_malloced_memory goes
4.6 -> 42.2 MiB at board load while current malloced_memory stays at 1.0.

Two theories tested and rejected rather than asserted:

  - 'Fewer loaded files means less compiler zone memory, so shipping less server JS
    pays twice.' Wrong. peak_malloced_memory is identical (42.1-42.5) across the
    preloadEntriesOnStart A/B despite 310 fewer files. The spike is a fixed cost of
    the board route, not a per-file one. Claim withdrawn.
  - Orphaned zlib streams. Next documents this leak class and the fix is absent from
    this build, and a peak snapshot showed 15 Gzip objects still present at idle -
    which read as accumulation. getActiveResourcesInfo() reports Zlib=0 both at
    baseline and after 300 deliberately aborted gzip requests, RSS flat. Snapshot
    objects are not live native handles, and nginx buffers responses so Next never
    sees the mid-response disconnect that orphans a stream.

Adds a /handles endpoint to the probe, which is what settled the second one.
2026-08-10 09:37:10 +02:00
Thomas Camlong e687c1dd22 docs(findings): full peak breakdown, the preloading correction, and jitless rejected
Records the Phase 2 peak work:

  - Corrects the earlier claim that Next preloading all 74 route graphs is
    'framework behaviour, not something this codebase sets'. Next documents
    experimental.preloadEntriesOnStart for exactly this, found by reading its
    memory-usage guide rather than by inference. Includes the deterministic
    per-stage table: 310 files and 6.3 MiB of source stay unloaded even after a
    full multi-tab session, so the cost is not deferred, it is never paid.
  - Rejects --jitless with evidence. Code space really does go 11.6 -> 0.0 MiB and
    peak RSS looked 24% better, but the flag was applied to the server while the
    browser's listener count fell 80% and board load got faster, which is
    impossible. The page under-rendered; the win was an app doing less work. The
    tell generalises: a server-side change moving a client-side metric means the
    workload changed and the comparison is void.
  - Notes that --optimize-for-size, --no-opt and --lite-mode are rejected outright
    by Node in NODE_OPTIONS, and that the rest of Next's memory guide covers
    build-time memory under Webpack, which does not apply to this Turbopack build.
2026-08-10 09:18:55 +02:00
Thomas Camlong bcea4733b9 perf(nextjs): stop preloading every route's module graph at startup
Next loads the module graph for all 74 app routes when the server starts. Measured
here: 1,114 files and 24 MiB of source resident before the first request, including
one shared 16 MiB subtree behind the API routes that reaches every integration
client. A dashboard serves a handful of routes per session.

`experimental.preloadEntriesOnStart: false` is Next's documented switch for this.
An earlier round of this work recorded route preloading as "framework behaviour, not
something this codebase sets" - that was wrong, and the findings doc is corrected.

Deterministic per-stage measurements, identical across both runs on each side:

  stage            source loaded      files        heapUsed
  boot idle       24.0 -> 10.6 MiB   1114 -> 706   70.7 -> 44.8  (-37%)
  post-login      27.9 -> 19.3 MiB   1215 -> 884   123.8 -> 109.8
  board loaded    29.0 -> 21.7 MiB   1323 -> 1006  (run-to-run noise)
  multi-tab       29.2 -> 22.9 MiB   1325 -> 1015  139.7 -> 127.8
  post-soak       29.2 -> 22.9 MiB   1325 -> 1015  142.5 -> 127.4

The important part is the last two rows: 310 files and 6.3 MiB of source stay
unloaded even after a full multi-tab session, so this is not merely deferred cost -
routes the session never visits are never paid for at all. Live heap is lower at
every stage, by 26 MiB at idle and 12-15 MiB under load and at rest.

Peak node RSS does not move (350.3 -> 346.1, spreads overlap). V8 sizes its heap
against available headroom, so a smaller live set does not shrink peak RSS; what
improves is idle footprint and how much room is left before the heap cap. Median
board load is +3% with overlapping spreads, so the first-hit cost this trades for
did not show at this workload.

Full E2E suite passes 8/8 against the built image.
2026-08-10 09:07:16 +02:00
Thomas Camlong 8594cc884f perf(docker): stop loading the gRPC stack dockerode only needs for image builds
dockerode/lib/docker.js required './session' at module scope, which pulls in
@grpc/grpc-js and @grpc/proto-loader - 112 files. Its single call site is inside
Docker.prototype.buildImage, the BuildKit path. Homarr lists, starts and stops
containers and never builds images, so every install with a Docker widget paid for
a gRPC stack it cannot reach.

Measured with a forced GC, in context rather than in a bare process: requiring
dockerode goes from +26.6 MiB RSS / 5.2 MiB heap to +20.6 MiB / 2.7 MiB, with grpc
and protobufjs files going 63 and 49 to zero.

The bare-process figure was +24.1 MiB, and quoting that would have overstated the
win by 4x - most of it overlaps with what dockerode loads anyway. The patch comment
records the in-context number.
2026-08-10 08:49:02 +02:00
Thomas Camlong 1de51a9afd docs(findings): record the DB-driver chop as measured, attempted, and backed out
mysql2 (77 files) and pg (13 files) load on every SQLite install because
db/drivers/index.ts imports all three wrappers statically. Re-measured with a
forced GC between readings, the prize is 21.4 MiB RSS - not the ~30 MiB the naive
sum implied, and drizzle-orm's 124.3 MiB in the require-cost sweep was almost
entirely that same no-GC artefact.

Three attempts, each caught by booting the real image rather than by typecheck:
createRequire(import.meta.url) is undefined in the CommonJS migrate.cjs bundle and
broke startup for every dialect; a dynamic require resolves to a .cjs build Next
never traces; and leaving drizzle's dialect static defeats the purpose because
drizzle-orm/mysql2 itself imports mysql2.

The two remaining routes are worse than the problem - a second copy of drizzle's
core alongside the ESM one, or hardcoding drizzle's internal file layout so an
upgrade breaks startup for MySQL and Postgres. Backed out; the structural blocker
is that createDb must stay synchronous.
2026-08-10 08:43:45 +02:00
Thomas Camlong 4a7da0bdab fix(benchmarks): split the allocation-site key explicitly instead of with replace
CodeQL flagged `key.replace("|", ...)` as js/incomplete-sanitization (high): it
replaces only the first occurrence. Here that was intentional - the key is
`functionName|url:line` and only the first separator matters - but written as a
replace it reads as a broken sanitiser, and it would mangle any URL containing a
pipe. Splitting on indexOf says what it means. Path trimming now uses replaceAll,
which is what was meant there too.
2026-08-09 22:24:46 +02:00
Thomas Camlong 3cc6ff33df perf(docker): stop loading ssh2, which allocates a 16 MiB WASM heap nobody uses
Profiling at the memory high-water mark - not at idle, which is what every earlier
round measured - found the largest single retainer in every snapshot was one V8
Context holding a 16.00 MiB ArrayBuffer. Expanding it through the dominator tree
named it: ssh2's crypto module, whose POLY1305_WASM_MODULE instantiates a
WebAssembly memory for ChaCha20-Poly1305 as soon as it is required.

Homarr has no SSH feature and does not depend on ssh2. The chain is
dockerode -> docker-modem -> ssh2, and docker-modem/lib/modem.js requires './ssh'
at module scope while using it in exactly one place, already guarded by
`protocol === 'ssh'`. A pnpm patch moves the require inside that branch. SSH-based
Docker hosts still work - require() is cached, so they pay it once - and every
socket-based install, which is everyone else, never loads it.

Measured over four alternated runs on a restored real backup, external memory:

                     before        after
  board loaded    25.4, 25.1    8.3, 11.2
  nav-01          28.5, 25.4   10.1,  8.6
  multi-tab       31.1, 35.8   13.0, 10.0
  post-soak       21.4, 21.5    5.5,  7.2

Consistently -15 to -18 MiB, ranges non-overlapping, and 30 fewer files loaded
(1470 -> 1440).

Peak RSS does NOT improve, and that is worth stating plainly: V8 sizes its heap
against available headroom, so freeing native memory lets it grow further before
collecting. What changes is the composition - a fixed 16 MiB of unusable WASM
buffer becomes headroom the app can actually use, which matters under a container
memory limit but does not move the peak number.

Also adds the instruments this needed. A heap snapshot can never explain a peak,
because V8 runs a full GC before serialising one and reports only survivors:

  - memory-probe.cjs gains in-process allocation sampling, on-demand snapshots via
    v8.writeHeapSnapshot, an smaps endpoint, and require-graph recording.
  - analyze-allocations.mjs diffs two stages to isolate what one stage allocated.
  - analyze-smaps-stages.mjs splits each stage's RSS into JS heap, JIT, native and
    file-backed pages.
  - analyze-heap-dominators.mjs can now expand the largest retainer, which is what
    identified ssh2. A plain graph walk escapes through __proto__ and is useless.

One trap recorded in the findings: taking the peak snapshot over CDP delivers it as
JS string chunks and added ~134 MiB of ArrayBuffers to the process being measured,
producing a fake 693 MiB reading. v8.writeHeapSnapshot streams from C++ instead, and
the probe now reports its own observerCostMiB so the cost of looking is visible.
2026-08-09 22:17:25 +02:00
Thomas Camlong 3bc25509cf docs(findings): record what the real Umami instance revealed about the page ceiling
The live site has >100k events for one event name, so the paged path logs that its
counts are a lower bound. The previous code had the identical 100k ceiling - the
46.3 MiB body was that ceiling being hit - but truncated in silence. The ceiling is
unchanged; it is now visible.
2026-08-09 21:42:13 +02:00
Thomas Camlong 4766653979 chore(benchmarks): record the server require graph, and use it to attribute boot-time code
A flat module cache says 24 MiB of source is resident before the first request, but not
what pulled any of it in. Recording parent -> child as the CommonJS loader resolves it
answers that, and the answer redirects the remaining 'ship less server JS' work:

  - Next loads all 74 app route bundles at startup, not lazily.
  - One 16.10 MiB / 460-file subtree is reached by the tRPC/API routes AND by
    instrumentation, so instrumentation's exclusive cost is 0 bytes - deferring it would
    free nothing, because Next preloads /api/[...trpc] anyway.
  - Only 1.10 MiB is common to every route, so this is the API surface pulling in every
    integration, not a universal barrel.
2026-08-09 21:39:28 +02:00
ajnart 3c56eab58f style: auto-fix lint and format 2026-08-09 19:36:22 +00:00
Thomas Camlong cb3a2a360e perf(umami): stop buffering 100k raw events per widget refresh
Profiling the container end to end found one allocation site responsible for
most of the peak: a single 46.3 MiB JSON body, assembled from 3,946 socket
chunks, inside undici's fullyReadBody. That function holds the chunk list and
the joined copy at the same instant, so a body of size N costs about 2N.

The source was two Umami calls asking for pageSize=100000, with
getMultiEventTimeSeriesAsync running them for every tracked event name
concurrently via Promise.all. getEventNamesAsync was downloading every raw
event of the last 30 days purely to collect a handful of distinct name strings.

  - getEventNamesAsync now reads the aggregated /metrics?type=event endpoint.
  - getWebsiteEventTimeSeriesAsync pages 5,000 records at a time, folding each
    page into the bucket map before requesting the next.
  - getMultiEventTimeSeriesAsync runs series sequentially.

Instances that return aggregated {x,y} buckets still answer in one request, an
instance that accepts pageSize but ignores page is detected rather than
double-counted, and hitting the page ceiling logs instead of silently
undercounting.

Alternated runs, 2 per side, on a restored real backup:

  peak container       498.9 -> 398.5 MiB  (-20%)
  peak node RSS        452.4 -> 353.3 MiB  (-22%)
  peak during 4 tabs   498.5 -> 356.3 MiB  (-29%)
  peak arrayBuffers     95.8 ->   6.6 MiB  (-93%)
  largest allocation    46.4 ->   2.4 MiB  (-95%)

Costs +4% CPU from the extra round trips; board load latency unchanged.
Post-soak steady state and boot idle are unchanged within run-to-run spread.

Adds the attribution tooling this needed: attribute-memory.mts reconciles
cgroup -> processes -> address-space mappings -> V8 spaces -> loaded code,
memory-probe.cjs is an injectable diagnostics preload (never shipped) that
tracks buffer and fetch allocation sites, analyze-heap-dominators.mjs computes
real retained sizes, and compare-stress-runs.mjs reports alternated runs with
their spread.

Also corrects an earlier claim in PHASE2-FINDINGS.md: the reported "7.86 MiB of
duplicate chunk sources" was an artefact of V8 truncating string contents to
~1 KiB in heap snapshots. There are zero byte-identical duplicates among all
3,055 server JS files. The real doubling is character width - one codepoint
above U+00FF forces a whole source string to two bytes per character, which
costs 9.45 MiB across 80 loaded files.

Adds 6 tests; this integration had none.
2026-08-09 21:35:00 +02:00
Thomas Camlong f707f64333 perf: externalise @ioredis/commands, and identify what actually holds the memory
Tested the hypothesis that cached base64 images were bloating TanStack Query and
Redis. They are not, measured on a loaded 4-tab session:

  base64 data-URLs, server heap   0.01 MiB (2 strings)
  base64 blobs, browser heap      0.01 MiB (7 strings)
  Redis total                     2.17 MB
  image bytes in Redis            none - image-proxy stores URL + headers only

What the server heap is actually made of (141 MiB used): 41.4 MiB is bundled
JavaScript source retained as strings, ~29% of the heap. The memory is the code, not
cached data — which is why every cache-tuning idea in the fix catalog measured at
zero, and why serverExternalPackages was the only lever that moved idle memory.

7.86 MiB of that is literal duplicate copies: the same module inlined into several
Turbopack chunks and retained once per chunk. @ioredis/commands' command table was in
10 chunks and resident 6 times (2.07 MiB wasted), so it is now external — ioredis is
already external and owns it as a direct dependency, so it stays traced.

Recorded as bundle hygiene, NOT a memory win: idle 89.7 -> 89.4 MiB is inside noise,
and board-loaded/peak deltas were smaller than the run-to-run spread (one "before" run
read 220.6 MiB against 169-173 for identical code).

got and @octokit/request-error duplicate just as badly but were deliberately left
bundled: they are only transitive dependencies, and a check inside the image showed
got is not traced into node_modules, so externalising it would have turned a bundled
module into a require that resolves to nothing — a possible MODULE_NOT_FOUND for zero
measured gain.

analyze-heap-strings.mjs buckets a snapshot's strings by what they are, which is what
made "37 MiB of strings" actionable instead of just alarming.

Verified: 28/28 widgets render, auth control passes, zero module-not-found in the
container log, typecheck clean. pnpm test 1794/1795 with two different Testcontainers
flakes across runs (pi-hole, sabnzbd); both pass 18/18 in isolation.
2026-08-09 16:07:26 +02:00
Thomas Camlong 933a54ead9 docs(perf): consolidated before/after, plus what is still unverified
All previous numbers were spread across rounds with different instruments and
different workloads. This is one instrument, one workload, alternated, 2 runs per
side, merge-base vs HEAD:

  server idle        141.9 -> 89.7 MiB   (-37%)
  server loaded      257.3 -> 173.0 MiB  (-33%)
  server peak        705.2 -> 568.4 MiB  (-19%)
  after idle soak    332.0 -> 232.5 MiB  (-30%)
  JS event listeners 10,126 -> 7,774     (-23%)
  React fibers       5,224 -> 3,588      (-31%)
  DOM nodes          9,175 -> 8,284      (-10%)
  browser JS heap    55.2 -> 51.1 MiB    (-7%)
  image              410 -> 398 MB       (-3%)
  container CPU      22.8 -> 22.8 s      unchanged

Recorded prominently rather than buried: under *sustained* 4-tab load the two builds
converge, 368 vs 364 MiB. The wins are idle, steady state and peak. That is why
hammering the app with five tabs did not feel 30% lighter — at saturation the
footprint is live widget data and render churn, which this PR does not change.

Also documents the gaps, since "tested" without them would be misleading: no
multi-hour soak (the #5301 reports are about days), single user, SQLite only, the
DevTools Profiler tab confirmed at bundle level rather than in the extension, and
the App Router re-rendering ~1x/second still unfixed.

verify-context-menu.mts now also covers touch long-press, and no longer asserts on a
fiber total — that counter swung 7,454 to 16,149 between runs of the same build, so
it was noise. probe-fiber-tree.mts is the stable counter and is what the -31% above
comes from.
2026-08-09 15:39:41 +02:00
Thomas Camlong 63e65bbc6c perf(common): stop useTimeAgo re-rendering every second, add re-render profiler
measure-rerenders.mts counts React commits on an untouched board and attributes each
to the shallowest component that re-rendered, i.e. where the update entered the tree.

What it found: an idle board commits 73-96 times per minute, re-rendering
@mantine/core/Box over 30,000 times in 60s. That is the churn behind the 650 MB
browser peak — transient allocation rather than retention. About 80% of those commits
enter at the same anonymous depth-2 provider, props [value, children], value {appDir}:
Next.js's internal HeadManagerContext, whose value is a fresh object literal on every
App Router render, so each render re-renders the whole app beneath it. The rest enter
at __next_metadata_boundary__ / <title>, pointing the same way.

The trigger is inside Next's router/metadata handling — application code has no
polling router.refresh() — so nothing here fixes it. It is the next thing to chase and
it dominates everything downstream of it.

useTimeAgo called setState on a 1s interval even when the label was unchanged ("3
minutes ago" holds for a full minute), re-rendering its subtree ~60x/minute per
consumer for an identical string. It now compares before setting. Recorded as a
correctness fix and NOT a measured win: idle commits stayed inside their 73-96 range
afterwards, because the app-root churn dominates.

pnpm test 1794/1795 (pre-existing pi-hole Testcontainers flake); typecheck clean
across 37 packages; no new lint errors.
2026-08-09 15:18:38 +02:00
ajnart 9bebc28102 style: auto-fix lint and format 2026-08-09 11:40:56 +00:00
Thomas Camlong 605431bc3f fix(perf): actually produce a React profiling build
DevTools still reported "Profiling not supported" because the build was never a
profiling one, despite three settings appearing to enable it. Next applies the React
swap through a webpack alias — getReactProfilingInProduction, mapping
'react-dom/client$' to 'react-dom/profiling' — that Turbopack never sees. A Turbopack
resolveAlias does not catch it either: the vendored entry requires
./cjs/react-dom-client.production.js relatively, not through a bare specifier.

The Dockerfile now performs that same swap on the vendored entry when
HOMARR_PROFILING=true, guarded so it fails the build loudly if React's file layout
changes rather than silently producing a normal build again.

Added `pnpm docker:profiling`, which builds and then verifies, because
`docker build . -t homarr:performance` alone yields a non-profiling image and that is
almost certainly what was being run.

verify-react-profiling.mts checks the bundle rather than the runtime: 2 client chunks
carry React's profiling renderer with the flag, 0 without. It replaces a Playwright
probe that could not tell the two apart — React sends a reduced inject payload to a
synthetic devtools hook, byte-identical for both builds, so that probe reported a
false negative for profiling and production alike. The bundle marker discriminates;
confirm the Profiler tab itself with the real extension.
2026-08-09 13:39:43 +02:00
ajnart 1d09172887 style: auto-fix lint and format 2026-08-09 09:06:26 +00:00
Thomas Camlong b4be884dfc perf(board): one shared context menu instead of one per widget
Mantine's Menu mounts Popover + PopoverDropdown + several context providers even
while closed, so a menu per widget was ~13 fibers each for something almost never
opened. WidgetContextMenuProvider now mounts a single menu for the board, anchored
to the cursor; items keep only a contextmenu/long-press listener.

The dropdown takes the section's gridstack ref as a prop rather than reading
SectionContext — the shared menu lives outside any section, and reading it from
context there crashed the board with "useSectionContext must be used within a
SectionContext".

Verified against both builds with scripts/benchmarks/verify-context-menu.mts
(opens on right-click, shows actions, closes on Escape, re-targets to a second
widget):

  idle fibers          18,670 -> 16,278  (-12.8%)
  fibers after use     +4,191 -> +4
  JS event listeners   10,128 -> 7,774   (-23%)
  DOM nodes             9,189 -> 8,287   (-10%)
  browser JS heap       54.1  -> 52.7 MiB
  median board load      520  -> 479 ms
  container CPU         20.4  -> 20.0 s

Two runs per side, four tabs across two boards. Server memory is unchanged, as
expected for a client-only change. The listener reduction is the point: the 650 MB
session had ~88k listeners and this removes a per-widget source of them.

The gate is validated by passing against unmodified code first. An earlier version
used Playwright's click({ button: "right" }), which emits no contextmenu event
here and so failed shipped code identically — it wrongly condemned a previous
attempt at this optimisation. A gate never run against the un-optimised build
proves only that something changed, not that it still works.

pnpm test 1794/1795 (pre-existing pi-hole Testcontainers flake, 9/9 in isolation);
typecheck clean across 37 packages; 0 lint errors.
2026-08-09 11:05:03 +02:00
Thomas Camlong fd7bcb2231 test(perf): fiber-tree probe, and the client-side diagnosis it produced
probe-fiber-tree.mts walks the live React tree and reports which components are
mounted, what DOM they render and their props. Heap snapshots only give minified
names like `y`, and source maps do not help because the snapshot records runtime
names — props shape plus rendered DOM identifies a component reliably. It injects
a DevTools hook stub before load so React registers every root; climbing from one
DOM node's fiber missed most of the tree (2,093 fibers vs 3,918).

What it found, on a 28-widget board:

- The most-mounted component is @mantine/core/Box (513). It is the primitive every
  Mantine component renders through, so it is a symptom of how much is mounted
  rather than a bug. Proportions match the 650 MB session snapshot (13% vs 17%),
  so the minified `y` there is almost certainly Box.
- 71 Menu -> Popover -> PopoverDropdown triplets are mounted for context menus that
  are almost never opened. Isolated with the existing enableRightClickOnWidgets
  flag: 3,918 fibers with them, 3,554 without — 364 fibers, 9%, scaling with item
  count. The dropdown contents are already lazy; the wrappers are not.

A fix that deferred the Menu until first right-click was reverted: it broke
right-click and inflated the tree to 20,107 fibers, because the replayed
contextmenu event re-entered the arming handler. Doing it properly needs one
board-level menu instead of one per item. The Playwright check that caught this is
in the findings — a fiber count alone would have read as a win.

Also records that the demo account had completed_board_tour = 0, so earlier rounds
measured the first-time-user path; completing it moves the tree by 38 fibers, so
the published numbers stand.
2026-08-09 10:37:33 +02:00
Thomas Camlong 8f18b17207 test(perf): profiling build, container introspection, client-side measurement
The container-only harness was measuring the wrong thing. Real usage — several
tabs, several boards, clicking into widgets — reaches ~880 MB server and a
650 MB *browser* heap, and the browser was never measured at all.

Profiling build: `docker build --build-arg HOMARR_PROFILING=true . -t
homarr:performance`. React's production bundle strips the hooks DevTools needs,
hence "Profiling not supported". Note reactProductionProfiling in next.config is
read only by the webpack path; under Turbopack it comes from `next build
--profile`, and turbo strips any env var not declared in the task's `env`, so all
three had to line up before the flag actually took effect — verified by markers
present only in React's profiling bundle, and static chunks going 20 -> 84 MB
with source maps.

inspect-container.mts attaches to an already-running container (SIGUSR1 opens
Node's inspector — no rebuild, no restart), reports cgroup/per-process/V8 memory
and can pull a heap snapshot out over CDP. The inspected context has no require
and no dynamic import, so it reads globals and uses CDP domains instead.

analyze-heapsnapshot / analyze-react-fibers / analyze-heap-retainers summarise a
snapshot too large for DevTools to open.

Harness now opens STRESS_BOARDS across STRESS_TABS_PER_BOARD tabs, clicks into
widgets, and records browser heap, DOM nodes and listener count via CDP.
Performance.getMetrics silently returns nothing unless the domain is enabled,
which is why the first run reported a 0 MiB client heap.
2026-08-09 08:21:57 +02:00
Thomas Camlong a532f9d377 docs(perf): repetition-backed headline, round-3 result, unpursued levers
Headline moves from single-sample -36% to repetition-backed -34% idle /
-35% loaded (2 before, 3 after, groups non-overlapping).

Round 3 (API-docs SPA lazy-loaded, MCP SDK external) is recorded as NOT a
memory win: every figure lands inside the previous build's spread, so it is
credited only with the deterministic -3 MB image.

Also records the measured cost of the levers not pursued: in-image Redis is
8.1 MiB RSS (the catalog guessed 10-40 MB) and nginx 9.2 MiB RSS / 3.1 PSS,
so neither justifies its deployment risk. next-server is 225.8 of ~236 MiB
at board-loaded, so anything further has to come out of that heap.
2026-08-09 03:21:21 +02:00
Thomas Camlong 61d9faffc5 perf: keep the API-docs SPA and MCP SDK out of the server graph
@scalar/api-reference-react is a documentation SPA reachable only from one
admin tab, and it renders nothing usable on the server, yet it was compiled
into the server graph for an SSR pass whose output is discarded. It is a
client component already, so next/dynamic with ssr: false is allowed in
place and needs no restructuring. @modelcontextprotocol/sdk is only reachable
through /api/mcp/[transport], so it moves to serverExternalPackages.

Honest accounting: this is an image-size and architecture change, not a
measured memory win. Idle landed at 89.4 MiB against 90.3/91.1/100.5 for the
previous build — the better end of that spread, but inside it, so on one
sample it cannot be claimed as a reduction. What is deterministic is image
size, 401 -> 398 MB. Kept because a docs SPA should not be server-rendered
regardless of what the memory number does.

pnpm typecheck and lint clean; pnpm test 1794/1795 with the known
Testcontainers pi-hole flake, which passes 9/9 in isolation.
2026-08-09 03:20:28 +02:00
Thomas Camlong 1965c75356 docs(perf): record round-2 findings and the auth proof
Adds the GC-tuning and wider-unbundling numbers, states the CPU and image-size
costs alongside the memory wins, and documents the three assertions that now
prove the measurement runs behind a real session rather than a public page.
2026-08-09 01:44:19 +02:00
Thomas Camlong 79ea987761 perf: bound V8's young generation and unbundle server-only deps
Two further wins, both measured on a restored 28-widget board behind a real
demo session (scripts/benchmarks/stress-restore.mts).

--max-semi-space-size=4 in run.sh. --max-old-space-size only bounds the old
generation; V8 also sizes the young generation generously (16 MB per
semi-space) and that slack is pure resident memory. 4 MB cut idle anonymous
memory 120.0 -> 97.2 MiB and peak 710.6 -> 560.2 MiB. It costs CPU, as any
GC-space reduction does — 15.8 -> 17.1 CPU-seconds over the run, ~8% — with
no board-load latency regression (median 594 -> 570 ms). Overridable via
HOMARR_MAX_SEMI_SPACE_SIZE.

serverExternalPackages for the remaining server-only dependencies. Bundling
a package inlines its transitive graph into a chunk that is compiled at boot
even when the code path never runs; leaving it external makes it a lazy
require. Adding winston, drizzle-orm, ical.js, jszip, ldapts, node-unifi,
@kubernetes/client-node and linkedom took idle to 90.3 MiB and board-loaded
209.6 -> 164.6 MiB, with slightly lower CPU (16.3s) and latency (549 ms).
The trade-off is image size: 388 -> 401 MB, because those packages now ship
in node_modules instead of being tree-shaken into chunks.

Cumulative vs the merge-base: idle 141.7 -> 90.3 MiB (-36%), board loaded
257.8 -> 164.6 MiB (-36%), under load 392.7 -> 302.5 MiB (-23%).

The harness now proves the session rather than inferring it: it asserts an
anonymous request to the private board is denied, requires the Auth.js
session cookie after sign-in, and reports whether it had to modify the
restored data at all (now "none"). It also records cgroup CPU seconds and
board-load latency, so a memory win bought with CPU cannot hide.

pnpm test 144/144 files, 1795 tests; typecheck and lint clean.
2026-08-09 01:43:06 +02:00
Thomas Camlong 9826000722 docs(perf): record post-trim verification
Removing the three zero-effect changes cost nothing: idle 120.0 MiB,
board loaded 229.5, after stress 357.0 — all within noise of the
untrimmed branch, with the lowest peak of any config measured.
2026-08-09 01:03:59 +02:00
Thomas Camlong b3f58cd8aa perf: drop the cache layers that measured at zero
Ablation (PHASE2-FINDINGS.md) attributed the memory win to the
serverExternalPackages line and the integrations barrel diet. These three
measured at exactly zero and were most of the file churn, so they go:

next-cache-handler: orphaned. cacheHandlers and cacheComponents are both
already gone from next.config.ts and nothing imported the package — it
was reachable only through a stale dependency entry.

cache-tags / cache-invalidation: there is no "use cache", cacheTag,
cacheLife or unstable_cache anywhere in app source, so revalidateTag was
firing on tags nothing registers, with no handler to receive them, inside
a catch that swallowed the resulting out-of-scope error. Removing the
calls also removed the SELECTs that existed only to supply a board name
or member list to them, so board.ts, group.ts, user.ts, serverSettings.ts,
section-router.ts and change-search-preferences.ts return to the
merge-base exactly. The live in-process integration data cache is
untouched.

Redis 6 -> 2 client consolidation: no measurable effect at idle (139.3 vs
139.2 MiB) or under load. Reverted, but the pubSub:last:* TTL is kept —
that one is an unbounded-growth fix, not a consolidation.

23 files, -883 lines. Diff vs merge-base: 100 files/3336 insertions ->
82 files/2484. pnpm test 144/144 files, 1795 tests, 0 failures;
typecheck and lint clean; app and Docker image build.
2026-08-09 00:59:41 +02:00
Thomas Camlong 84dcbd8847 docs(perf): attribute the Phase 1 memory win to specific changes
Measured on a populated instance (real backup restored through the
onboarding UI, 28-widget board open) rather than an empty container, and
compared on anonymous memory because memory.current swung 192-261 MiB
for the same image once page cache was counted.

Branch vs merge-base: -24% idle, -11% board loaded, -8% under load.

Cumulative ablation from the baseline image shows where that comes from:
adding mysql2/pg to serverExternalPackages — one line — accounts for the
entire idle win (-21 MiB), because the driver barrel statically imports
all three drivers and bundling inlined mysql2's whole graph into the
boot-loaded server chunks (108 MB of chunks vs 95 MB, and mysql2 absent
from node_modules because it had been inlined). The integrations barrel
diet carries the under-load win but moves idle by 0. Redis 6->2 clients
and the V8 heap cap both move idle by 0 — the cap is a drift ceiling,
not a reduction, since removing it entirely (cap4096) matched the branch
at every stage.

The cache-invalidation layer and next-cache-handler package measure at
exactly zero: no "use cache", cacheTag, cacheLife or unstable_cache
anywhere in app source, and cacheHandlers/cacheComponents are already
gone from next.config.ts, so 66 revalidateTag call sites across ~30
routers are no-ops. Keep/drop decision left open.
2026-08-09 00:44:50 +02:00
Thomas Camlong fb80b3882d test(benchmarks): compare on anon memory, retry pre-hydration login clicks
Two fixes found by running the harness against a real pair of images:

Login submitted before the form hydrated, so the click was a no-op that
looked identical to bad credentials. Load the page at networkidle and
retry the submit, surfacing any page alert if all attempts fail — the
same race already handled for the onboarding restore toggle.

memory.current counts reclaimable page cache, which moved the "same"
baseline between 192 and 261 MiB across runs. Anonymous memory is now
the headline metric (idle was 144 MiB anon vs 77 MiB of page cache), so
an A/B is not decided by unrelated disk I/O.
2026-08-09 00:44:50 +02:00
Thomas Camlong 6363d96a1b test(benchmarks): fix post-login wait, allow V8 flag injection
Sign-in redirects client-side, so page.waitForURL never reached its
implicit "load" state and every run failed after the restore step even
though authentication had succeeded. Poll window.location instead.

STRESS_NODE_OPTIONS injects NODE_OPTIONS directly, which is what makes
the V8 heap cap attributable: the merge-base run.sh does not read
HOMARR_MAX_OLD_SPACE_SIZE, so the cap can only be A/B'd against that
image by passing the flag in.
2026-08-09 00:44:50 +02:00
Thomas Camlong 0d19968131 test(benchmarks): add populated-dashboard memory stress benchmark
Container memory was only ever measured on an empty install, which hides
the cost that actually matters: a real board with real widgets and
integrations loaded.

stress-restore.mts boots a fresh container, restores a supplied backup
through the onboarding restore UI (including the "I understand"
confirmation), signs in, opens the busiest board, waits for every widget
to mount, then hammers navigation and soaks idle while sampling cgroup
v2 memory at each stage.

The backup is passed by path via STRESS_BACKUP_ZIP and never enters the
repo — it carries real integration secrets and an encryption key.
.gitignore now also refuses homarr-backup*.zip and loose *.sqlite.

Baseline reading (merge-base image, 28-item board, all widgets ready):
223 MiB idle -> 448 MiB board loaded -> 688 MiB peak.
2026-08-09 00:44:50 +02:00
ajnart 9187af8882 style: auto-fix lint and format 2026-08-08 11:02:04 +00:00
Thomas Camlong abec79abe8 fix: address CodeRabbit review — typed mocks, robust benchmark script
- Replace `as any` casts in integration-test-connection.spec.ts with
  Pick<Integration, "testConnectionAsync"> + mockResolvedValue
- build-compare.sh: split local declarations (SC2318), clean .next
  between builds with --force, add EXIT trap for safe stash/cleanup
2026-08-08 13:00:42 +02:00
Thomas Camlong b1713a9820 fix: revert Suspense layout wrappers causing E2E hydration errors
Revert layout.tsx, _layout-creator.tsx, manage/layout.tsx,
_theme.tsx, _creator.tsx, mantine.tsx, and login/page.tsx
to origin/dev — the Suspense provider splits caused
"EditMode is required", stream close, and React #419 errors.

Also remove cacheComponents/cacheHandlers from next.config.ts
(incompatible with auth/i18n/dynamic rendering) and fix
Boolean() truthy bug in Redis connection.
2026-08-08 12:40:48 +02:00
Thomas Camlong 3db87d781f fix: address CodeRabbit review issues
- redis-handler: add TLS_CA and USERNAME env support matching core client
- connection.ts: fix Boolean() truthy bug (CI="false" was disabling Redis)
- channel.ts: add in-memory lock for single-instance fallback mode
- channel-subscription-tracker: lazy-init subscriber client
- board.ts: invalidate user cache on board delete for home-board users
- .env.example: document REDIS_CACHE_DB vs DATABASE_INDEX collision risk
- CLI docs: fix docker exec command (remove /bin/bash prefix)
- Docker docs: fix docker rm -f and docker ps -a for stopped containers
2026-08-08 09:17:39 +02:00
Thomas Camlong 51467ded12 docs: remove obsolete X-Frame-Options ALLOW-FROM examples
ALLOW-FROM is not supported by any modern browser. The CSP
frame-ancestors directive is the correct replacement and was
already present in both examples.
2026-08-08 09:09:53 +02:00
Thomas Camlong 38fc9bfd01 chore: remove full-screen loading spinners from boards and manage 2026-08-08 09:09:19 +02:00
Thomas Camlong 9bc1fc397c chore: remove .cursor/goal files from git 2026-08-08 09:08:48 +02:00
Thomas Camlong 9596a83f60 fix: use any cast for mock integration in test-connection spec
The barrel no longer exports PiHoleIntegrationV6 after the dynamic
import refactor. The mock only needs testConnectionAsync, so a type
cast is sufficient.
2026-08-08 09:03:18 +02:00
Thomas Camlong e036cd86d8 perf: move dependencies to build-time static JSON, add build benchmark
- Move getDependenciesAsync filesystem scan to scripts/dump-dependencies.mts
  (runs at build time, outputs static-data/dependencies.json)
- About page now imports dependencies.json like other static data
  (eliminates runtime glob of all package.json files)
- Delete dead getDependenciesAsync + parseDependencyCatalogsAsync from
  package-reader.ts (only getPackageVersion remains)
- Add scripts/benchmarks/build-compare.sh for dev vs branch comparison
2026-08-08 08:58:54 +02:00
Thomas Camlong 6692acbd33 chore: strip formatting noise, delete dead routes, fix CodeRabbit issues
- Remove 179 format-only file changes (docs, migrations, plex, patchmon, CSS)
- Delete .cursor/goal files from git tracking
- Delete 4 dead /api/about routes (unused, data loaded from static JSON)
- Fix resolveAlias to use absolute paths (CodeRabbit)
- Fix Tailwind classes in Mantine app layout fallback (CodeRabbit)
- Return stale data during cache revalidation (CodeRabbit)
2026-08-08 08:45:16 +02:00
Thomas Camlong 95a7d56989 fix: pi-hole test import from barrel to direct path
The barrel no longer exports PiHoleIntegrationV5/V6 after the
dynamic integration import refactor.
2026-08-08 00:41:50 +02:00
Thomas Camlong 7bd5991076 ci: trigger CI run 2026-08-08 00:29:48 +02:00
Thomas Camlong 345a4be4b0 fix: CI build errors + simplify cache handler and data cache
- Remove `export const dynamic` from 4 routes (incompatible with cacheComponents)
- Static import for redis-handler (Turbopack can't resolve relative dynamic imports)
- Simplify redis-handler: Buffer.concat, NonNullable, single connect promise
- Inline single-use helpers in integration-data-cache
- Non-greedy SVG regex to prevent ReDoS
- Inline getDevelopmentServiceAliases in next.config
- Remove restating comments
2026-08-08 00:19:27 +02:00
Thomas Camlong 6f563640ac fix: remove as const from svgSanitizeOptions to fix DOMPurify type compat 2026-08-07 22:12:09 +02:00
Thomas Camlong 4270f2f2f1 fix: address CodeRabbit review — stream tee, cache guard, base64, layout fallback
- BoardLayoutFallback: use HomarrLogoWithTitle (no board context required)
- integration-data-cache: guard cache.set with placeholder identity to
  prevent stale write-back after invalidation + test
- next-cache-handler: memoize the promise (not resolved handler) to
  prevent concurrent Redis connections
- redis-handler: tee ReadableStream before consuming in set(), hand fresh
  branch to each get() caller
- redis-handler: base64-encode body instead of number[] (3-4x smaller)
2026-08-07 22:09:11 +02:00
ajnart 00855d4825 style: auto-fix lint and format 2026-08-07 19:44:19 +00:00
Thomas Camlong 39334569d6 fix: remove notifications from refetch-intervals to match widget definitions
The notifications widget has no refetchInterval in its definition,
so the static polling policy entry was causing a manifest test failure.
2026-08-07 21:42:08 +02:00
Thomas Camlong ffed73ba8f merge: incorporate latest dev changes into phase1-perf branch
Merge origin/dev to pick up service isolation (concurrently dev script,
turbopack aliases, instrumentation simplification), PostHog analytics,
color scheme, and other dev improvements while preserving all Phase 1
performance work (Suspense boundaries, cache components, Redis handler,
integration barrel diet, cache invalidation).
2026-08-07 21:36:48 +02:00
Thomas Camlong 90d9a50227 fix: correct Content-Type for content-detected SVG responses
When looksLikeSvg detects SVG content under a non-SVG MIME type,
set Content-Type to image/svg+xml so browsers render correctly
with X-Content-Type-Options: nosniff.
2026-08-07 21:25:07 +02:00
Thomas Camlong e836d07d4c fix: content-based SVG detection, guard tests, document REDIS_CACHE_DB
- Detect SVG by content (magic bytes) in addition to MIME type to prevent
  sanitization bypass via mislabeled uploads
- Add guard test verifying DOMPurify actually strips dirty payloads
- Add looksLikeSvg detection test
- Document REDIS_CACHE_DB in .env.example
2026-08-07 21:22:25 +02:00
Thomas Camlong 33293ba1cf refactor: simplify Phase 1 infrastructure code
- cache-invalidation: extract isExpectedRevalidateError helper
- integration-data-cache: flatten early-return chain, extract eviction
- redis-handler: extract StoredEntry type, readStreamBody, ensureConnected
- svg-purify: single getOwnPropertyDescriptor lookup per prop
- session-scoped-providers: simplify auth flow
- retryable-loader: use function declarations
- translation/request: replace spread-in-reduce with for loop
2026-08-07 21:04:42 +02:00
Thomas Camlong b3b2d5a022 fix: SVG regex prefers </svg> terminator, add self-closing children test
Use greedy <svg>...</svg> match preferring closing tag over inner />
to prevent truncation at first self-closing child element.
Self-closing root <svg .../> handled as separate alternative.
Added test with <path/> + <circle/> children asserting well-formed output.
2026-08-07 21:00:55 +02:00
Thomas Camlong 67ac2bbc7c fix: Redis expire guard, transferOwnership prev-owner, SVG regex
- Redis get() rejects legacy expire=0 entries and cleans them up
- transferOwnership invalidates previous owner in addition to new
- SVG regex uses non-greedy match and handles self-closing <svg />
- SVG fallback returns empty string instead of HTML wrapper
- Memory channel comment: FIFO eviction, not timestamp-based
2026-08-07 20:57:07 +02:00
Thomas Camlong fbb2eddd87 fix: concurrent stale race, Redis expire leak, missing invalidation, UI Suspense
- Fix integration data cache: set pending synchronously before creating
  fetch to prevent concurrent stale-revalidation from spawning duplicate
  upstream calls
- Redis cache handler: skip persisting entries with expire=0 (dynamic
  entries should not be stored in Redis)
- Add missing invalidation: user.changePassword, group.updateGroup,
  group.transferOwnership
- Move SearchEngineOptimization inside Suspense in root layout
- Add Suspense fallback loader to root layout
- Move searchParams resolution inside Suspense in login page
2026-08-07 20:53:36 +02:00
Thomas Camlong 38ac8f3075 fix: concurrent cache miss, SVG output, narrowed error handling, format
- Fix integration data cache: check pending promise before TTL to
  prevent returning undefined on concurrent cold-cache access
- Add concurrent-miss test (two parallel callers, one fetch)
- SVG sanitizer: extract <svg> from DOMPurify output to strip
  residual HTML wrapper
- Narrow safeRevalidateTag catch: log unexpected errors, silence only
  known "outside request scope" messages
- Run oxfmt across all packages (206 files formatted)
2026-08-07 20:48:17 +02:00
Thomas Camlong b81b90875f fix: revert DB driver to static imports, fix test regressions
- Revert createRequire() for DB drivers — require() can't resolve .ts
  files in Vitest, breaking 49 test suites
- Make cache invalidation fail-safe via try-catch around revalidateTag
  (throws outside Next.js context: tests, tasks service, websocket)
- Fix 6 integration test imports: barrel no longer exports concrete
  classes, point tests at direct source paths
- Fix Redis cache handler lint errors: remove non-null assertions,
  suppress __NEXT_BUILD_ID underscore warning
2026-08-07 20:43:07 +02:00
Thomas Camlong 67fcd69eb8 fix: inverted tag expiration in Redis handler + saveItemOptions board name
- Fix critical bug: tag expiration comparison was inverted (< should be >)
  in Redis cache handler, which would prevent revalidateTag from clearing
  stale entries
- Pass board name to invalidateBoardCache in saveItemOptions
- Add clarifying comment about two-tier invalidation helpers
2026-08-07 19:30:58 +02:00
Thomas Camlong aeabfd3521 fix: requestMedia dual-layer cache invalidation
Mirror answerRequest pattern: invalidate both module-level integration
data cache and request-handler caches for media request list/stats.
2026-08-07 19:27:41 +02:00
Thomas Camlong 37930e44cc perf: clean integration barrel — remove 46 concrete class re-exports
- Remove all concrete integration class re-exports from barrel
- Replace instanceof check with kind check in health-monitoring
- Document cache invalidation as preparatory (cacheTag follow-up)
2026-08-07 19:22:32 +02:00
Thomas Camlong 451dc3ce39 fix: address review #4 — board name invalidation, SVG tests, test gaps
- Pass board.name to all invalidateBoardCache calls (permission mutations)
- Align integration create invalidation with update/delete
- Add initial-fetch-failure test for integration-data-cache
- Expand SVG XSS test corpus (javascript:, xlink:href, @import, onbegin)
- Forbid <style> tags in SVG sanitizer to prevent CSS injection
- Fix duplicate export in packages/common/src/index.ts
2026-08-07 19:18:18 +02:00
Thomas Camlong 4361c8dbe2 fix: address Bugbot review #3 — mutation invalidation gaps
- Dual-layer cache invalidation on interact mutations
  (dns-hole, media-requests, smart-home, indexer-manager, anchor-notes)
- importOldmarrConfig: invalidate created boards
- createBoard: invalidateUserCache when auto-setting homeBoardId
- group.savePositions: invalidate user caches for affected members
- integration.requestMedia: invalidate integration data cache
- user.delete: invalidate user cache
2026-08-07 19:14:02 +02:00
Thomas Camlong 0f8362c3da fix: address Bugbot review #2 findings
- Invalidate integration data cache from widget interact mutations
  (downloads, dns-hole, media-requests, indexer-manager, smart-home, anchor-notes)
- Redis handler getExpiration: fetch missing tags from Redis (multi-instance coherence)
- saveItemOptions: invalidate board cache after saving
- Remove unused cacheTags.app
- Fix integration data cache comment accuracy
2026-08-07 19:08:57 +02:00
Thomas Camlong 6b45aca5d5 fix: address Bugbot review findings
- Skip module-level integration cache for multi-instance (REDIS_IS_EXTERNAL)
- Add invalidateBoardCache to permission mutations (prevents future leak)
- Remove unused cacheTags.integrationData
- Replace KEYS with SCAN in Redis cache handler refreshTags()
- Document REDIS_CACHE_DB env var
2026-08-07 19:04:53 +02:00
Thomas Camlong 018f98e318 perf: implement Phase 1 homepage diet via Next.js Cache Components
P1-0: V8 heap cap (HOMARR_MAX_OLD_SPACE_SIZE, default 512MB)
P1-1: Module graph cuts
  - Dynamic integration imports (55 static → lazy loaders)
  - Conditional DB driver loading (only configured driver loaded)
  - Replace jsdom with linkedom for SVG sanitization
  - Delete eager widget barrel (55 import * as → 0)
  - Extract createRetryableLoader to @homarr/common
P1-2: Enable cacheComponents + restructure root layout
  - Suspense boundaries around session-dependent providers
  - ColorSchemeScript replaces server-side cookie reads
  - Board/manage/auth layouts wrapped in Suspense
P1-3: Cache read paths
  - Cache tag registry + invalidation helpers
  - Cache invalidation wired into all mutation routers
  - Translation deep-merge memoized per locale
P1-4: Prerendered shells + loading.tsx files
P1-5: Shared integration data cache (60s TTL, stale-while-revalidate)
P1-6: Redis-backed Next.js cache handler (@homarr/next-cache-handler)
P1-7: Redis optional (memory channel fallback), TTL on pubSub:last:*,
      Redis clients consolidated from 6 to 2
2026-08-07 18:59:47 +02:00
Thomas Camlong 366ec6bdbc fix custom widget e2e login race 2026-08-03 15:52:08 +02:00
Thomas Camlong 9f4b147f0e fix legacy media server option mapping 2026-08-03 15:29:27 +02:00
Thomas Camlong 08af704c8a Merge origin/dev into feat/build-performance 2026-08-03 15:21:23 +02:00
Thomas Camlong 3f9fb215c6 fix: avoid ambiguous benchmark regex 2026-08-03 14:27:21 +02:00
Thomas Camlong de801f3f00 perf: optimize dashboard loading and builds 2026-08-02 15:58:16 +02:00
ajnart d90e762e4b fix: align Navidrome media sessions 2026-07-31 13:52:39 +02:00
ajnart 81a13d3b52 Merge remote-tracking branch 'origin/dev' into feat/build-performance 2026-07-31 13:38:16 +02:00
ajnart 388a74086e fix: address Copilot cache review 2026-07-31 13:32:48 +02:00
ajnart cd6c83a7b0 test: guard LDAP user lookup 2026-07-30 00:50:02 +02:00
ajnart 9963ba0274 fix: harden runtime isolation and regression tests 2026-07-30 00:38:34 +02:00
ajnart 5d9f05e083 test(e2e): open spotlight with hotkey 2026-07-29 07:53:35 +02:00
ajnart d48f299585 test(e2e): wait for authenticated board 2026-07-29 07:42:57 +02:00
ajnart c80e9878de fix: handle lazy startup failures 2026-07-29 07:23:09 +02:00
ajnart 84f87ce24e fix(ci): bound affected task concurrency 2026-07-29 07:11:26 +02:00
ajnart fbb57c8f40 fix(docker): build runtime workspaces only 2026-07-29 06:57:06 +02:00
ajnart 0b9c4d6375 fix: address performance review feedback 2026-07-29 06:50:13 +02:00
ajnart 7266e15356 style: auto-fix lint and format 2026-07-29 06:47:07 +02:00
ajnart c22c733463 build(docker): simplify manifest copy layers 2026-07-29 06:47:07 +02:00
ajnart 440626ba00 test(performance): add authenticated local benchmarks 2026-07-29 06:47:07 +02:00
ajnart f8d8a09c56 perf(proxy): isolate settings database reads 2026-07-29 06:47:07 +02:00
ajnart add6ee56ba fix(updates): enforce a persisted daily check 2026-07-29 06:47:07 +02:00
ajnart 6b74078766 perf(widgets): lazy load the application graph 2026-07-29 06:47:07 +02:00
ajnart 4ca0b747bb chore(next): upgrade preview and isolate dev services 2026-07-29 06:47:07 +02:00
2753 changed files with 77706 additions and 254574 deletions

No files matched your search

-40
View File
@@ -1,40 +0,0 @@
---
name: codebase-context
description: Navigate Homarr's monorepo architecture and reuse shared packages. Use when locating code, choosing a package boundary, understanding runtime services, tRPC, databases, widgets, integrations, cron jobs, Custom Widgets, Workshop, or checking @homarr/common before creating a utility.
---
# Homarr Codebase Context
Orient from the current checkout before editing. Treat package manifests, exports, and source as authoritative when they differ from this snapshot.
## Workflow
1. Confirm the branch and inspect the nearest `package.json`, existing sibling modules, and package exports.
2. Search `@homarr/common` before creating a general-purpose helper. Read [references/common.md](references/common.md) for every public entrypoint and export.
3. Read [references/architecture.md](references/architecture.md) when the task crosses packages, runtimes, routing, data, widgets, integrations, cron jobs, Custom Widgets, or Workshop.
4. Keep the change in the deepest package that owns the behavior. Reuse public entrypoints rather than deep-importing implementation files.
5. Re-run targeted searches before documenting counts, consumers, or dependency edges; these change faster than the architectural seams.
## Current toolchain
- Use Node `24.18.0` from `mise.toml` and pnpm `11.15.1` from `package.json`.
- Use pnpm workspaces with Turborepo. Workspace packages use the `@homarr/` scope and catalog-managed dependencies.
- Use Next.js App Router, TypeScript, tRPC, Drizzle, Mantine v9, Tabler icons, Jotai, TanStack Query, next-intl, Vitest, and Playwright.
- Use oxlint and oxfmt. Mantine is the application UI system; Tailwind is limited to the docs app.
- `pnpm dev` starts only `@homarr/nextjs`. Start a standalone runtime explicitly when the task needs it.
## High-value seams
- Server/RSC tRPC: `@homarr/api/server`
- Client tRPC hooks: `@homarr/api/client`
- Main router: `packages/api/src/root.ts`
- MCP eager router: `packages/api/src/mcp.ts`
- Database schemas: `packages/db/schema/{sqlite,mysql,postgresql}.ts`
- Widget registry: `packages/widgets/src/registry.ts`
- Widget loading manifest: `packages/widgets/src/manifest.ts`
- Integration definitions: `packages/definitions/src/integration.ts`
- Integration factory: `packages/integrations/src/base/creator.ts`
- Cron jobs: `packages/cron-jobs/src/jobs/`
- App routes: `apps/nextjs/src/app/[locale]/`
When a code change affects users, invoke `documentation-sync`. When exposing tRPC through MCP, invoke `mcp-integration`.
@@ -1,4 +0,0 @@
interface:
display_name: "Homarr Codebase Context"
short_description: "Navigate Homarr architecture and shared APIs"
default_prompt: "Use $codebase-context to locate the right Homarr package and reuse existing shared APIs."
@@ -1,110 +0,0 @@
# Homarr architecture
## Contents
- [Applications and runtimes](#applications-and-runtimes)
- [Package map](#package-map)
- [Dependency direction](#dependency-direction)
- [Core patterns](#core-patterns)
- [Commands and conventions](#commands-and-conventions)
## Applications and runtimes
| Path | Role |
| ---------------- | ---------------------------------------------------------------------------------------- |
| `apps/nextjs` | Main Next.js application and HTTP APIs on port 3000 |
| `apps/docs` | Docusaurus documentation site on port 3003 |
| `apps/tasks` | Cron-job initialization and scheduling runtime; it does not expose the old port-3002 API |
| `apps/websocket` | tRPC WebSocket subscriptions on port 3001 |
| `apps/workshop` | Go/PocketBase Workshop service and provider integration |
`pnpm dev` runs only `@homarr/nextjs`. Use `pnpm --filter @homarr/tasks dev`, `pnpm --filter @homarr/websocket dev`, or `pnpm dev:docs` when a task needs those runtimes.
In production, `apps/nextjs/src/instrumentation-node.ts` starts the tasks and WebSocket runtimes inside the Node process. Nginx listens on port 7575, proxies `/websockets` to port 3001, and sends other traffic to port 3000. Redis can run internally or externally.
## Package map
### Foundation and domain
- `core`: environment validation, database and Redis factories, logging, infrastructure.
- `definitions`: domain enums, integration definitions, widget kinds, permissions.
- `common`: shared utilities with environment-specific entrypoints. Search it before adding helpers.
- `validation`: shared Zod schemas.
- `server-settings`: typed server-setting keys and access.
### Data, identity, and transport
- `db`: Drizzle schemas, queries, migrations, and three database drivers.
- `auth`: Auth.js/NextAuth configuration, sessions, providers, and API keys.
- `api`: tRPC routers, contexts, MCP metadata extraction, OpenAPI, and client/server entrypoints.
- `redis`: Redis clients, pub/sub channels, and caches.
- `cron-jobs-core`, `cron-jobs`, `cron-job-status`: scheduling primitives, job implementations, and status transport.
### Backends and external systems
- `integrations`: typed clients for external applications.
- `request-handler`: shared server request flows.
- `docker`, `ping`, `image-proxy`, `icons`, `analytics`: bounded infrastructure capabilities.
- `custom-widgets`: Custom JSX v2 schema, validation, runtime, and security contracts.
- `workshop`: TypeScript Workshop client, backend contracts, and schemas used by Homarr.
### UI and composed features
- `ui`: shared Mantine components, theme, and hooks. Search here before creating a primitive.
- `form`, `forms-collection`: Zod-backed form setup and reusable forms.
- `modals`, `modals-collection`: modal primitives and feature modals.
- `boards`: board state, edit mode, and cache coordination.
- `settings`, `spotlight`, `notifications`, `onboarding`: cross-feature UI flows.
- `widgets`: widget definitions, dynamic component loaders, and shared widget UI.
- `translation`: next-intl configuration and locale catalogs.
- `cli`: Node operations CLI. `tools/homarr-dev` is the separate Go developer CLI.
## Dependency direction
Prefer these directions; verify the exact dependency in package manifests before changing a boundary.
1. Put infrastructure in `core` and stable domain names in `definitions`.
2. Put reusable, dependency-light helpers in `common`; use its public environment-specific entrypoints.
3. Build persistence in `db`, identity in `auth`, transport in `redis`, and external clients in `integrations`.
4. Compose server behavior in `api`, `cron-jobs`, and request-handler packages.
5. Build primitives in `ui`, `form`, and `modals`; compose product surfaces in widgets and collection packages.
6. Keep app entrypoints in `apps/*` thin. Avoid moving reusable behavior into `apps/nextjs` when a package already owns it.
## Core patterns
### tRPC and MCP
`packages/api/src/root.ts` lazily composes the application router. Server components use `@homarr/api/server`; client components use `@homarr/api/client`. MCP extraction requires the eager router in `packages/api/src/mcp.ts`, so MCP-enabled subrouters must be imported and registered there directly. Use the `mcp-integration` skill for exposure and security rules.
### Databases
Drizzle schemas live in parallel files under `packages/db/schema/`. `DB_DRIVER` selects `better-sqlite3`, `mysql2`, or `node-postgres`; migrations live in matching directories under `packages/db/migrations/`. Apply schema behavior consistently across all supported drivers unless the task is explicitly driver-specific.
### Widgets
`packages/widgets/src/registry.ts` owns the explicit widget-module and component-loader registries. `packages/widgets/src/manifest.ts` caches those loaders and exposes definitions, components, and combined resources. Widget modules use `createWidgetDefinition` and typed option builders. Data commonly comes from tRPC queries or Redis-backed cron results. Use existing shared widget components before creating local variants.
### Integrations
`packages/definitions/src/integration.ts` owns `integrationDefs` and `IntegrationKind`. `packages/integrations/src/base/creator.ts` owns `createIntegrationAsync`; concrete integrations implement the shared base contract. Keep credentials and permission checks on server-owned paths.
### Cron jobs
Jobs live in `packages/cron-jobs/src/jobs/` and use `createCronJob`. `apps/tasks` initializes and starts the job group. Job results and status can flow through Redis and tRPC; the retired port-3002 management-API model is no longer part of the architecture.
### Routing and localization
User routes live below `apps/nextjs/src/app/[locale]/`; the locale segment does not add a URL prefix. Route groups organize home and board surfaces. Administrative routes live under `manage`, Custom Widget configuration has a token route, and Workshop has a management surface.
### Custom Widgets and Workshop
`packages/custom-widgets` owns the Custom JSX v2 contract. `packages/workshop` is the Homarr-side client/contract package; `apps/workshop` is the separate Go/PocketBase service. Use the `homarr-custom-widget` skill for authoring and security constraints.
## Commands and conventions
- Read current scripts from the root or package `package.json`; avoid caching commands that do not exist.
- Use `pnpm dev` for Next.js, `pnpm dev:docs` for docs, and `pnpm dev:cli -- <args>` for the Go developer CLI.
- Use `pnpm docker:dev:up` for the Redis development dependency.
- Use package-filtered typechecks or focused tests for the touched behavior.
- Use `~/*` only inside the Next.js app. Use public `@homarr/*` entrypoints across package boundaries.
- Keep Mantine application styles separate from the Tailwind-based docs app.
@@ -1,85 +0,0 @@
# `@homarr/common` public API
## Contents
- [Reuse workflow](#reuse-workflow)
- [Entrypoints](#entrypoints)
- [Root exports](#root-exports)
- [Types exports](#types-exports)
- [Server exports](#server-exports)
- [Client and environment exports](#client-and-environment-exports)
## Reuse workflow
Before writing a utility, search both the public name and the behavior:
```bash
rg -n "export (const|function|class|type|interface)|export \*" packages/common
rg -n "@homarr/common" apps packages
```
Import only from an exported package entrypoint. Do not deep-import `@homarr/common/src/*`. Check `packages/common/package.json` and the source again when adding or renaming an export; this inventory describes the current repository tree.
## Entrypoints
| Import | Source | Environment |
| ----------------------- | ---------------------------- | ----------------------------------------------------------------------------------- |
| `@homarr/common` | `index.ts` -> `src/index.ts` | Shared root; includes React hooks and some Next/Mantine-aware helpers |
| `@homarr/common/types` | `src/types.ts` | Type-only utilities |
| `@homarr/common/server` | `src/server.ts` | Node/server-only security, encryption, request, error, and bounded-response helpers |
| `@homarr/common/client` | `src/client.ts` | Client-facing Next server action export |
| `@homarr/common/env` | `env.ts` | Validated package environment |
## Root exports
| Area | Public exports | Purpose |
| ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------- |
| Objects | `objectKeys`, `objectEntries`, `isRecord`, `hashObjectBase64` | Typed key/entry iteration, record guards, and stable query-key hashing |
| Strings | `capitalize`, `isNullOrWhitespace`, `bestMatch`, `normalizeImageName`, `getImageMatchRank` | Text normalization and matching |
| Cookies | `parseCookies`, `setClientCookie` | Cookie parsing and browser cookie serialization |
| Arrays | `splitToNChunks`, `splitToChunksWithNItems` | Split by number of groups or maximum group size |
| Dates | `formatDuration`, `toValidDate`, `isDateWithin` | Duration formatting, safe date coercion, relative-window checks |
| Timing | `Stopwatch` | Elapsed-time measurement and formatting |
| React hooks | `useTimeAgo`, `useIntegrationConnected` | Relative time and integration freshness state |
| URLs | `removeTrailingSlash`, `extractBaseUrlFromHeaders`, `getPortFromUrl`, `isAbsoluteUrl`, `SAFE_NEW_TAB_REL`, `getSafeApplicationUrl`, `getSafeAppHref` | URL normalization, safe navigation, ports, and new-tab policy |
| Numbers and bytes | `formatNumber`, `formatBytes`, `formatBytesPair`, `formatByteRate`, `formatBitRate`, `defaultByteUnitSystem`, `metricToImperial` | Compact numbers, byte displays, byte/bit rates, and unit conversion |
| Byte types | `ByteUnitSystem`, `FormatBytesOptions`, `FormatBitRateOptions` | Options for byte and bit-rate formatting |
| Errors | `extractErrorMessage`, `FlattenError` | Unknown-error messages and flattenable domain errors |
| Theme | `getMantineColor` | Resolve a Mantine color and shade |
| Functions | `isFunction` | Unknown-value function guard |
| Grid | `GridAlgorithmItem`, `generateResponsiveGridFor` | Responsive nested-section grid placement |
| IDs | `createId` | CUID2 ID creation |
The root deliberately does not re-export `types`, `server`, `client`, or `env`; use their named entrypoints.
## Types exports
`@homarr/common/types` exports:
- `MaybePromise<T>`
- `AtLeastOneOf<T>`
- `Modify<T, R>`
- `RemoveReadonly<T>`
- `MaybeArray<T>`
- `Inverse<T>`
- `inferSearchParamsFromSchema<TSchema>`
## Server exports
`@homarr/common/server` exports:
- Security: `generateSecureRandomToken`.
- Encryption: `encryptSecret`, `decryptSecret`, `decryptSecretWithKey`, `createKeyedFingerprint`, `verifyKeyedFingerprint`.
- Requests: `userAgent`, `DeviceType`, `ipAddressFromHeaders`.
- Bounded bodies: `readBoundedJsonResponseAsync`.
- Parse errors: `ParseError`, `ParseErrorHandler`, `ZodParseErrorHandler`, `JsonParseErrorHandler`.
- HTTP handlers: `HttpErrorHandler`, `FetchHttpErrorHandler`, `matchErrorCode`, `OFetchHttpErrorHandler`, `AxiosHttpErrorHandler`, `TsdavHttpErrorHandler`, `OctokitHttpErrorHandler`.
- HTTP errors: `RequestError`, `ResponseError`, `requestErrorMap`.
- HTTP error types: `AnyRequestError`, `AnyRequestErrorInput`, `RequestErrorInput`, `RequestErrorType`, `RequestErrorReason`, `AnyRequestErrorReason`, `RequestErrorCode`.
`NodeFetchHttpErrorHandler` exists as an implementation file but is not exported by the public server entrypoint. Add an intentional public export before using it across packages.
## Client and environment exports
- `@homarr/common/client` exports `revalidatePathActionAsync`, a Next server action for page revalidation.
- `@homarr/common/env` exports `env` with validated `NODE_ENV`, `SECRET_ENCRYPTION_KEY`, and `NO_EXTERNAL_CONNECTION` values.
@@ -1,51 +0,0 @@
---
name: documentation-sync
description: Keep Homarr's Fumadocs documentation aligned with user-facing code. Use when adding or changing integrations, widgets, APIs, environment variables, CLI commands, authentication, permissions, cron jobs, settings, or UI behavior that users must understand.
---
# Documentation Sync
Update `apps/docs/` in the same change as user-facing behavior. Inspect adjacent pages and the docs types before copying a pattern; the source and docs evolve together.
## Map the change
| Code change | Documentation target |
| ----------------------------------- | -------------------------------------------------------------------- |
| New integration | `apps/docs/docs/integrations/<slug>/index.mdx` and `index.ts` |
| New widget | `apps/docs/docs/widgets/<slug>/index.mdx` and `index.ts` |
| Changed public or management API | `apps/docs/docs/management/api/index.mdx` |
| New or changed environment variable | Relevant page under `apps/docs/docs/advanced/` or installation docs |
| New or changed CLI command | `apps/docs/docs/advanced/command-line/` |
| Authentication/provider change | Relevant SSO or authentication page under `apps/docs/docs/advanced/` |
| UI or workflow change | Relevant getting-started, management, widget, or integration page |
| Cron-job behavior | `apps/docs/docs/management/tasks.mdx` |
| Permission behavior | `apps/docs/docs/management/users.mdx` and the affected feature page |
| Custom Widget or Workshop behavior | Relevant Custom Widget, Assistant, or Workshop documentation |
If no existing row fits, search by the user-visible term and update the page where a user would look for the behavior:
```bash
rg -n "<term>" apps/docs/docs apps/docs/src
```
## Follow existing typed patterns
For an integration:
- Add or update `apps/docs/docs/integrations/<slug>/index.ts` with the local `IntegrationDefinition` pattern.
- Add or update `index.mdx` with the established `IntegrationHeader`, `IntegrationCapabilites`, and `IntegrationSecrets` components as applicable.
- Reuse metadata from `@homarr/definitions` when the docs app already exposes it.
For a widget:
- Add or update `apps/docs/docs/widgets/<slug>/index.ts` with the local `WidgetDefinition` pattern.
- Add or update `index.mdx` with the established `WidgetHeader`, `WidgetConfig`, and `AddingWidget` components as applicable.
- Document options, required integrations, permissions, empty states, and changed behavior that affect setup or use.
## Completion criteria
1. Describe the shipped behavior, not an implementation plan.
2. Keep names, defaults, paths, screenshots, links, and prerequisites consistent with code.
3. Update every affected page and remove superseded guidance.
4. Run the narrowest useful docs validation. Use `pnpm turbo build --filter=@homarr/docs` when links, MDX, generated definitions, or navigation can fail; otherwise run the docs package formatter on touched files.
5. Treat broken links and anchors as failures; run `pnpm --filter @homarr/docs validate:links` after the static build.
@@ -1,4 +0,0 @@
interface:
display_name: "Homarr Documentation Sync"
short_description: "Keep user-facing docs aligned with code"
default_prompt: "Use $documentation-sync to map this Homarr code change to the required documentation updates."
@@ -1,43 +0,0 @@
---
name: homarr-custom-widget
description: Author, validate, preview, test, install, or configure API-backed Homarr Custom JSX v2 widgets.
---
# Homarr Custom Widget
Author one widget or a coordinated set of widgets. Load only needed release-matched context. Run lifecycle tools alone; independent reads may run together. For a set, research once and finish each widget's validation, evidence, and persistence before the next.
1. Read primary API documentation. Use web search when documentation is not supplied or may have changed.
2. Create credential-free definitions with keyed `sources`, `requests`, optional `options`, and safe JSX `template`.
3. While drafting, use `customWidget_validateTemplate` for focused JSX diagnostics without resending the manifest.
4. Send the definition once to `customWidget_previewCreate` and test its queries/actions. For a JSX-only fix, validate, call `customWidget_previewReviseTemplate` with its session, and retest; it inherits the manifest and resets evidence. Create a preview only for source/request/option changes.
5. Configure deployment-specific source URLs and credentials through Homarr; never repeat plaintext.
6. Persist each exact final tested preview with `customWidget_createFromPreview`. Do not resend a large definition through `customWidget_create` when a preview session is available.
Treat a supplied sample or successful preview response as the binding contract. Render every core requested field, guard optional arrays and nested values before indexing, and do not silently drop returned items. Humanize numeric enums with indexed literal label arrays, omit absent numeric values instead of inventing zero, and label timestamp timezones. Give recoverable load errors and empty states a clear refresh or retry path.
Use `{option:name}` or `$option` for saved options. Use `{param:name}` or `$param` for values supplied by `SubFetch`, `ActionButton`, or `ToggleSwitch`. Load queries cannot use invocation parameters. Render load queries from `data` and `status` with `RefreshButton`; reserve `SubFetch` for manual parameterized queries. Templates read `data`, `status`, `options`, and temporary `inputs`.
For a bound control that depends on another input, declare its default and set `resetKey` to that scalar dependency. For example, `<Pagination bind="page" defaultValue={1} resetKey={inputs.search} />` restores page 1 when a search changes without fetching by itself.
`SubFetch` with `trigger="manual"` renders its own load button. Pass a card or image through `triggerContent` with `triggerAriaLabel` when that content should launch the request. Its callback is `(result, meta)`; callback names must not shadow the reserved roots `data`, `status`, `options`, or `inputs`. Never author `onClick` or a fetch callback. Use `Icon` or `TablerIcon` with a `name`; never invent an `IconFoo` component.
Plan capabilities. Use one `customWidget_findComponents` search per job. Batch selected non-obvious binding or interaction documentation with `customWidget_getComponents`; reserve `customWidget_getComponent` for one unknown-prop repair. Failed validation reopens discovery; otherwise search only for a missing capability. The full catalog is for broad exploration; load at most one example. Context never limits composition. Prefer clear hierarchy, responsive grids, divided lists, and aligned actions over nested row cards.
Do not simplify a useful workflow merely because the template is interpreted. Freely compose any supported installed components with multiple sources, requests, options, `choicesFrom`, bound filters, charts, responsive detail areas, manual queries, and safe actions when they improve the user's job. Complexity must remain purposeful rather than decorative.
Polish with a divided list or responsive media grid. Make one metric/action asymmetrically primary; keep headers compact, identity/state clear, metadata quiet, and one badge. At base artwork fills its row and caps above xs; never combine full-width media and nowrap.
Make initial states actionable and wrap variable labels/values on narrow tiles. Do not use an unlabeled decorative icon as an empty state.
Do not use imports, hooks, refs, raw HTML, raw event callbacks, browser requests, arbitrary functions, eval, bigint, npm packages, authored statement blocks, IIFEs, or recursion. Do not pretend MCP tools exist in an offline chat session.
For a community widget, use `customWidget_workshopSearch`, `customWidget_workshopGet`, then `customWidget_workshopInstall`; configure its source securely. Preview configuration expires, so create before persisting configuration.
Reference routing:
- Load `schema` once before a new manifest; otherwise only for concrete request or option ambiguity.
- Load `runtime` for bound inputs, manual queries, actions, or other interaction.
- Load `security` once for any authenticated source or mutation, or to resolve a URL or interpreter limitation.
Repository installations expose these as files under `references/`. MCP clients can call `customWidget_getReference` or read `homarr://custom-widgets/references/{name}` instead of loading every reference.
@@ -1,4 +0,0 @@
interface:
display_name: "Homarr Custom Widget"
short_description: "Build and deploy safe Custom JSX widgets"
default_prompt: "Use $homarr-custom-widget to author, validate, preview, and configure a safe Custom JSX v2 widget."
@@ -1,37 +0,0 @@
# Runtime
Templates read `data.requestId`, `status.requestId`, `options.name`, and temporary `inputs.name`. Status is `{ loading, ok, status, statusText, error }`. Render load queries directly from `data` and `status` with `RefreshButton`; never wrap them in `SubFetch`.
`bind="search"` creates an in-memory input. It is never persisted. Supply invocation values only through `params`, for example:
```jsx
<TextInput bind="search" label="Search" />
<Pagination bind="page" resetKey={inputs.search} defaultValue={1} total={5} />
<SubFetch requestId="search" trigger="manual" params={{ query: inputs.search }}>
{(result) => <Stack>{(result.results ?? []).map(item => <Text key={item.id}>{item.name}</Text>)}</Stack>}
</SubFetch>
```
Manual queries require `trigger: "manual"` on request and `SubFetch`; otherwise they run automatically. `triggerContent` with `triggerAriaLabel` makes custom content the launcher. `SubFetch` owns loading/error/retry; its child receives success plus `{ ok, status, statusText, loading: false }`. Never author `onClick` or fetch callbacks.
`SubFetch`, `ActionButton`, and `ToggleSwitch` need literal `requestId`; validation rejects missing/computed IDs.
Inside a successful manual result, `<RefreshButton requestId="search" label="Run again" />` reruns the same parameters.
When a manual SubFetch request ID, parameters, or effective definition changes, Homarr immediately hides its prior result and returns to the trigger. It cannot fetch the new parameters until the user triggers it again.
The `SubFetch` callback receives the entire JSON response exactly as previewed. If the response is `{ "results": [...] }`, render and map `result.results`; never map the envelope itself. Trace every rendered field from the preview response before persistence.
Format timestamps with safe static helpers; never use `new Date`. Never invent a formatter component. Use `Date.toLocaleString(value, "en-US", "UTC")` plus a visible `UTC` label. Also available: `Date.toISOString`, `Date.toLocaleDateString`, and `Date.toLocaleTimeString`.
For compact numeric enums, index a literal label array with a fallback:
```jsx
<Text>{["Unknown", "Pending", "Ready"][(item.status ?? 1) - 1] ?? "Unknown"}</Text>
```
Every stateful control must use `bind`, and its `inputs.<name>` value must feed a supported request/helper when it is meant to change remote data. For dependent pagination, declare `defaultValue={1}` and use `resetKey={inputs.search}` to restore page 1 when the query changes. If a control cannot affect the workflow through a binding, option, or runtime helper, render concise context instead of a dead control.
Callback parameters must not shadow the reserved roots `data`, `status`, `options`, or `inputs`. Use `<Icon name="refresh" />` or `<TablerIcon name="refresh" />`; never invent components such as `<IconFoo />`.
Use expression callbacks for supported collections and trusted slots. No callback blocks, IIFEs, authored recursion, or raw events. Regex is limited to safe string operations.
@@ -1,49 +0,0 @@
# Schema
```ts
interface HomarrCustomWidgetV2 {
$schema: "homarr-custom-widget-v2";
name: string;
description?: string;
iconUrl?: string;
sources: Record<string, CustomWidgetSource>;
requests: Record<string, CustomWidgetRequest>;
options?: Record<string, CustomWidgetOption>;
template: string;
}
```
The object key `default` is the required source ID; `default` is not a property on a source. Source properties are `name?`, `baseUrl`, `networkScope`, and `auth?`:
```json
{
"sources": {
"default": {
"name": "Service",
"baseUrl": "http://service.local:5055/api/v1",
"networkScope": "private",
"auth": { "type": "apiKeyHeader", "name": "X-Api-Key" }
}
},
"requests": {
"summary": { "path": "/summary" },
"search": { "trigger": "manual", "path": "/search", "query": { "q": { "$param": "query" } } },
"create": {
"kind": "action",
"method": "POST",
"path": "/items",
"body": { "id": { "$param": "id" } },
"confirmation": "Create this item?",
"invalidates": ["search"]
}
}
}
```
Auth is `none`, `bearer`, `basic`, `{ "type": "apiKeyHeader", "name": "X-Api-Key" }`, or `{ "type": "apiKeyQuery", "name": "api_key" }`. A request defaults to source `default`, kind `query`, method `GET`, trigger `load`, inherited auth, and view permission. Set `trigger: "manual"` for a parameterized query. An action defaults to manual and modify permission. DELETE uses full permission and confirmation. Do not use `load: false`.
Use stable real URLs for public APIs and clear suggested URLs for self-hosted services. Homarr collects the installer's server URL, network scope, and credentials as source setup; credentials remain outside the manifest.
Paths use `{option:name}` and `{param:name}`; query/body references use `{ "$option": "name" }` and `{ "$param": "name" }`. Constants stay primitive (`take: 10`); `$param` is only for manual helpers, never load queries. Names and types are inferred.
Every option has `label`, `control`, and `default`. Optional fields are `description`, `choices`, `choicesFrom`, `min`, `max`, `step`, `advanced`, and `group`.
@@ -1,7 +0,0 @@
# Security
All requests use Homarr's protected server executor. Source origin, network scope, DNS, redirects, SSRF, rate limits, permissions, size limits, timeouts, and encrypted credential injection remain enforced.
The JSX interpreter blocks imports, hooks, refs, raw event callbacks, browser requests, eval, arbitrary functions, prototype access, unsafe URLs, global CSS escape, arbitrary portals, bigint, statement blocks, IIFEs, and recursion. Regex literals must be bounded and reject backreferences, lookbehind, nested quantifiers, excessive length, and unsupported flags.
Credentials are stored separately and never exported or returned to an agent. A published self-hosted source URL is only a suggestion: installers must confirm or replace private and loopback URLs for their own Homarr deployment. Source origins cannot be controlled through widget options.
-98
View File
@@ -1,98 +0,0 @@
---
name: mcp-integration
description: Expose Homarr tRPC procedures safely through MCP. Use when adding or changing tRPC procedures under packages/api, adding MCP metadata, registering an eager MCP router, shaping tool inputs and descriptions, or reviewing MCP authorization, sensitive data, and destructive actions.
---
# MCP Integration
Expose a procedure only when an AI client should call it. Preserve the procedure's existing authorization and add MCP-specific safeguards for sensitive or destructive behavior.
## Assess exposure
Prefer:
- Data queries: list, search, get, health, stats, and summaries.
- User-requested actions with explicit permission checks and bounded inputs.
- Serializable results that do not reveal credentials or internal-only state.
Skip subscriptions, session/onboarding internals, file streams, blobs, credential values, and procedures whose authorization or audit behavior is unclear. Ask for a security review when a mutation is destructive, tenant boundaries are ambiguous, or returned data is sensitive.
## Add metadata
Place `.meta()` before `.input()` and `.query()` or `.mutation()`:
```typescript
getAll: protectedProcedure
.meta({
mcp: {
enabled: true,
description: "List resources the current user can access, including stable IDs used by resource_get",
},
})
.input(z.object({ limit: z.number().int().min(1).max(100).default(50) }))
.query(async ({ ctx, input }) => {
// Preserve normal procedure authorization.
});
```
No-input procedures can omit `.input()`; the MCP extractor accepts an empty object for them.
For mutations, state the effect and required permission in the description:
```typescript
remove: permissionRequiredProcedure
.meta({
mcp: {
enabled: true,
description: "Delete one resource by ID. Requires full access; get the ID from resource_all",
},
})
.input(z.object({ id: z.string() }))
.mutation(async ({ ctx, input }) => {
// Perform the permission-scoped action.
});
```
## Write discoverable descriptions
Include every item that applies:
- What the tool returns or changes.
- Supported integrations or services.
- Required IDs and the tool that returns them.
- Permission requirements and the meaning of permission fields.
- Important bounds, confirmation semantics, or irreversible effects.
Use precise domain language. A description such as `Get calendar events` is insufficient because it omits supported services, required IDs, and result scope.
## Shape inputs for clients
- Use a top-level `z.object({...})` for parameterized tools.
- Bound strings, arrays, numeric ranges, pagination, and payload sizes.
- Use defaults where a safe, unsurprising value exists.
- Prefer explicit fields over a top-level union or discriminated union.
- Keep identifiers stable and describe how to obtain them.
- Keep secrets out of inputs unless the procedure is explicitly a credential-configuration flow with server-owned encryption and non-return guarantees.
## Register the eager router
The application router in `packages/api/src/root.ts` is lazy. MCP tool extraction is synchronous, so add an eager import and registration in `packages/api/src/mcp.ts`:
```typescript
import { resourceRouter } from "./router/resource";
export const mcpRouter = createTRPCRouter({
resource: resourceRouter,
});
```
Register the smallest router that owns the enabled procedures. If a parent router uses `lazy()`, import the needed subrouter directly. Metadata on a procedure is not enough; a procedure absent from `mcpRouter` is not exposed.
## Validate the tool surface
1. Confirm the procedure still enforces its ordinary auth and permission boundary.
2. Confirm `packages/api/src/mcp.ts` eagerly includes the procedure.
3. Confirm the generated tool name, description, and JSON schema are unambiguous.
4. Run the focused API MCP spec when the tool list or extraction changes: `pnpm test packages/api/src/test/mcp.spec.ts`.
5. Run the focused route spec when protocol or transport behavior changes: `pnpm test apps/nextjs/src/app/api/mcp/[transport]/route.spec.ts`.
6. If manually probing `/api/mcp/<transport>`, use a scoped API key and avoid printing credentials or secret-bearing results.
@@ -1,4 +0,0 @@
interface:
display_name: "Homarr MCP Integration"
short_description: "Expose tRPC procedures safely through MCP"
default_prompt: "Use $mcp-integration to expose this Homarr tRPC procedure as a safe, discoverable MCP tool."
-1
View File
@@ -1 +0,0 @@
../.agents/skills
+206
View File
@@ -0,0 +1,206 @@
---
description: Comprehensive Homarr codebase architecture and conventions reference
alwaysApply: true
---
# Homarr Codebase Context
Homarr is an open-source, self-hosted dashboard for managing homelab services. It integrates with 50+ self-hosted apps (media servers, download clients, DNS, NAS, etc.) and provides a drag-and-drop widget-based UI.
## Tech Stack
- **Monorepo**: pnpm workspaces + Turborepo
- **Framework**: Next.js (App Router, `output: "standalone"`) — T3 Stack
- **Language**: TypeScript throughout
- **ORM**: Drizzle (supports SQLite via better-sqlite3, MySQL, PostgreSQL)
- **API**: tRPC (HTTP + WebSocket subscriptions) with superjson + OpenAPI bridge via `trpc-to-openapi`
- **Auth**: NextAuth v5 (database sessions, Credentials/LDAP/OIDC providers)
- **UI**: Mantine v9 (not Tailwind) + Tabler icons
- **State**: Jotai (atoms), TanStack Query (server state via tRPC)
- **Realtime**: WebSocket server (`ws`) for tRPC subscriptions, backed by Redis pub/sub
- **Cron**: `node-cron` in a standalone Fastify service (`apps/tasks`)
- **i18n**: next-intl with `[locale]` dynamic segment (prefix mode: `"never"`, locale from cookie)
- **Testing**: Vitest + jsdom, Playwright for E2E
- **Lint/Format**: oxlint + oxfmt (not ESLint/Prettier)
- **Docs**: Docusaurus 3 in `apps/docs/` (`@homarr/docs`)
- **Package manager**: pnpm 10.34.1, Node >= 24.16.0
## Repository Structure
```
homarr/
├── apps/
│ ├── nextjs/ # Main Next.js application (port 3000)
│ ├── docs/ # Docusaurus 3 documentation site (@homarr/docs)
│ ├── tasks/ # Cron job runner + Fastify tRPC API (port 3002)
│ └── websocket/ # Standalone tRPC WebSocket server (port 3001)
├── packages/
│ ├── api/ # tRPC appRouter, procedures, OpenAPI
│ ├── auth/ # NextAuth config, providers, session, API keys
│ ├── db/ # Drizzle schema (3 DB drivers), migrations, queries
│ ├── core/ # Env validation, DB/Redis driver factories, logging
│ ├── definitions/ # Domain enums: WidgetKind, IntegrationKind, permissions
│ ├── widgets/ # All 39 dashboard widgets (definitions + components)
│ ├── integrations/ # Integration classes (HTTP clients to external apps)
│ ├── redis/ # Redis pub/sub channels, caching abstractions
│ ├── translation/ # next-intl setup, locale configs, lang JSON files
│ ├── ui/ # Shared Mantine components, theme, hooks
│ ├── validation/ # Shared zod schemas for API/forms
│ ├── common/ # Shared utilities, IDs, errors
│ ├── cron-jobs/ # Cron job implementations (25+ jobs)
│ ├── cron-jobs-core/ # Cron scheduling primitives
│ ├── cron-job-api/ # tRPC router for job management (start/stop/trigger)
│ ├── cron-job-status/ # Cron status via Redis
│ ├── boards/ # Board context, edit mode, cache updater
│ ├── modals/ # Modal primitives on Mantine
│ ├── modals-collection/ # Feature modals (apps, boards, docker, etc.)
│ ├── form/ # useZodForm (Mantine + zod resolver)
│ ├── forms-collection/# Reusable form UIs (new app, icon picker, upload)
│ ├── spotlight/ # Command palette / search with multiple modes
│ ├── request-handler/ # Server request handlers (feeds, integrations)
│ ├── notifications/ # Mantine notifications wrapper
│ ├── docker/ # Dockerode-based Docker access
│ ├── icons/ # Icon DB/repo integration
│ ├── image-proxy/ # Image proxy + caching
│ ├── ping/ # Reachability / ping utilities
│ ├── analytics/ # Server-side analytics (Umami)
│ ├── server-settings/ # Server setting keys/types
│ ├── settings/ # User-facing settings UI context
│ ├── old-import/ # Legacy Homarr import
│ ├── old-schema/ # Legacy Homarr zod schemas
│ └── cli/ # Node CLI for ops (brocli)
├── tooling/
│ ├── typescript/ # Base tsconfig
│ └── github/ # CI setup action
├── development/ # Dev docker-compose (Redis, MySQL, PostgreSQL)
├── e2e/ # E2E test specs
└── Dockerfile # Multi-stage production build
```
## Three Runtime Services
| Service | Port | Role |
|---------|------|------|
| Next.js | 3000 | Main web app (SSR + API routes) |
| WebSocket | 3001 | tRPC subscriptions via `ws` |
| Tasks | 3002 | Cron job runner + management API via Fastify |
In Docker, nginx listens on **7575** and proxies to all three. The tasks service authenticates via `CRON_JOB_API_KEY` header.
## Package Dependency Layers
1. **Foundation**: `core` (env, DB drivers, Redis client, logging) → no `@homarr` deps
2. **Domain**: `definitions` → `common` → `core`
3. **Data**: `db` → `core`, `definitions`, `common`
4. **Auth**: `auth` → `db`, `definitions`, `core`, `validation`
5. **Cross-cutting**: `translation`, `validation`, `redis`, `server-settings`
6. **Feature backends**: `integrations`, `request-handler`, `cron-jobs`, `docker`, `ping`
7. **API surface**: `api` → pulls in auth, db, redis, integrations, etc.
8. **UI stack**: `ui`, `modals`, `form`, `notifications`, `boards`, `spotlight`
9. **Composed features**: `widgets`, `modals-collection`, `forms-collection`
## Key Patterns
### tRPC
- **Server caller** (RSC): `import { api } from "@homarr/api/server"` → cached caller with auth context
- **Client hooks**: `import { clientApi } from "@homarr/api/client"` → React Query hooks
- **Router structure**: `packages/api/src/root.ts` defines top-level namespaces (`user`, `board`, `widget`, `integration`, `docker`, `kubernetes`, etc.)
- **Procedures**: `publicProcedure`, `protectedProcedure`, `permissionRequiredProcedure`, `onboardingProcedure`
- **WebSocket subscriptions**: Client uses `wsLink` when `type === "subscription"`, connects to port 3001
### Multi-Database Support
Drizzle schemas exist in three parallel implementations:
- `packages/db/schema/sqlite.ts`
- `packages/db/schema/mysql.ts`
- `packages/db/schema/postgresql.ts`
Selected at runtime via `DB_DRIVER` env (`better-sqlite3` | `mysql2` | `node-postgres`). Migrations are per-engine in `packages/db/migrations/{sqlite,mysql,postgresql}/`.
### Widget System
Widgets are registered in `packages/widgets/src/index.tsx` as `widgetImports` (must satisfy `Record<WidgetKind, ...>`). Each widget:
- Has a `definition` created via `createWidgetDefinition(kind, { icon, createOptions, supportedIntegrations? })`
- Uses `optionsBuilder.from(factory => ({ ... }))` for typed options
- Loads dynamically via `next/dynamic`
- Fetches data via tRPC client hooks (backed by cron job caches)
### Integration System
Integrations are defined in `packages/definitions/src/integration.ts` (`integrationDefs`) and instantiated via `createIntegrationAsync` in `packages/integrations/src/base/creator.ts`. Each extends abstract `Integration` class with `testConnectionAsync` and typed methods.
### Cron Job System
Jobs defined in `packages/cron-jobs/src/jobs/`. Each job:
- Uses `createCronJob` with a cron expression
- Publishes results to Redis channels
- Widget subscriptions consume those channels via tRPC subscriptions
- Job management (start/stop/trigger) via `cronJobApi` HTTP client to port 3002
### Modals
Created with `createModal(component).withOptions({ defaultTitle })` from `@homarr/modals`. Opened via `useModalAction(SomeModal)`. Feature modals live in `packages/modals-collection/`.
### Forms
`useZodForm` from `@homarr/form` wraps Mantine `useForm` with zod validation. Shared form UIs in `packages/forms-collection/`.
### Routing (App Router)
- `apps/nextjs/src/app/[locale]/` — all UI routes under locale segment
- Route groups: `(home)`, `(content)`, `(board)` — don't affect URLs
- Key routes: `boards/[name]`, `manage/...` (admin), `auth/login`, `init` (onboarding), `widgets/[kind]`
- Locale driven by cookie (no URL prefix)
### Auth Providers
Configured via `AUTH_PROVIDERS` env. Supports: `credentials` (local), `ldap`, `oidc`. Database sessions with httpOnly cookies. API key auth for programmatic access.
## Environment Variables
Key env vars (from `.env.example`):
- `DB_DRIVER`: `better-sqlite3` | `mysql2` | `node-postgres`
- `DB_URL`: Database connection string
- `AUTH_SECRET`: NextAuth secret
- `SECRET_ENCRYPTION_KEY`: AES-256-CBC for integration secrets
- `AUTH_PROVIDERS`: Comma-separated (`credentials`, `ldap`, `oidc`)
- `CRON_JOB_API_KEY`: Shared key between Next.js and tasks service
- `REDIS_*`: Redis connection (host, port, password)
- `ENABLE_KUBERNETES`: Optional Kubernetes integration
## Development Commands
- `pnpm dev` — runs all three services in parallel via Turborepo
- `pnpm docker:dev` — starts Redis + MySQL + PostgreSQL containers
- `pnpm db:push` — push schema to SQLite (dev)
- `pnpm db:studio` — Drizzle Studio
- `pnpm test` — Vitest unit tests
- `pnpm test:e2e` — Playwright E2E
- `pnpm lint` / `pnpm format` — oxlint + oxfmt
- `pnpm dev:docs` — Docusaurus docs site (port 3003)
- `pnpm typecheck` — tsc --noEmit across all packages
## Conventions
- Workspace packages use `@homarr/` scope
- All versions managed via pnpm `catalog:` in workspace
- Path alias `~/*` → `./src/*` in Next.js app
- Mantine for all UI (no Tailwind) — primary color is red, `autoContrast: true`
- Icons from `@tabler/icons-react`
- Drag-and-drop via `@dnd-kit/*`
- `typescript.ignoreBuildErrors: true` in next.config (types checked separately via `typecheck`)
- `serverExternalPackages`: `dockerode`, `isomorphic-dompurify`, `jsdom`
## Documentation (`apps/docs`)
Docusaurus 3 site at `apps/docs/`, package name `@homarr/docs`.
- Run with `pnpm dev:docs` from root
- Build with `pnpm turbo build --filter=@homarr/docs`
- Uses Tailwind CSS 3 for custom pages (not Mantine)
- Imports `@homarr/definitions` for type-safe integration/widget metadata
- When adding or modifying widgets, integrations, cron jobs, API routes, env vars, or user-facing features, you MUST also update corresponding docs in `apps/docs/docs/`
- Integration docs: `apps/docs/docs/integrations/<slug>/index.mdx` + `index.ts`
- Widget docs: `apps/docs/docs/widgets/<slug>/index.mdx` + `index.ts`
- Strict link checking: broken links/anchors cause build failure
+50
View File
@@ -0,0 +1,50 @@
---
description: Keep documentation in sync with code changes
globs:
- packages/definitions/src/**
- packages/widgets/src/**
- packages/integrations/src/**
- packages/api/src/**
- apps/nextjs/src/**
- apps/docs/**
---
# Documentation Sync Rule
When modifying code that affects user-facing behavior, you MUST also update the documentation in `apps/docs/`.
## What triggers a doc update
| Code change | Doc location |
|---|---|
| New integration | `apps/docs/docs/integrations/<slug>/index.mdx` + `index.ts` |
| New widget | `apps/docs/docs/widgets/<slug>/index.mdx` + `index.ts` |
| Changed API surface | `apps/docs/docs/management/api.mdx` |
| New/changed env vars | `apps/docs/docs/advanced/` |
| New CLI commands | `apps/docs/docs/advanced/command-line/` |
| Changed auth providers | `apps/docs/docs/advanced/` SSO pages |
| UI/UX changes | Relevant getting-started or management pages |
| New cron job | `apps/docs/docs/management/tasks.mdx` |
| Changed permissions | `apps/docs/docs/management/users.mdx` |
## Integration doc pattern
Each integration has a folder under `apps/docs/docs/integrations/<slug>/`:
- `index.ts` — exports a typed `IntegrationDefinition` with name, description, iconUrl, path
- `index.mdx` — uses shared components: `IntegrationHeader`, `IntegrationCapabilites`, `IntegrationSecrets`
- Import metadata from `@homarr/definitions` where possible
## Widget doc pattern
Each widget has a folder under `apps/docs/docs/widgets/<slug>/`:
- `index.ts` — exports a typed `WidgetDefinition` with icon, name, description, path, configuration
- `index.mdx` — uses shared components: `WidgetHeader`, `WidgetConfig`, `WidgetAdding`
## Development
- Run docs locally: `pnpm dev:docs` from root
- Build docs: `pnpm turbo build --filter=@homarr/docs`
- The docs app can import from `@homarr/definitions` for type-safe integration/widget metadata
- Docusaurus strict mode throws on broken links — fix any broken references before committing
@@ -0,0 +1,622 @@
# Migrating Tables from mantine-react-table to mantine-datatable
This document captures every pattern, fix, and decision from the downloads widget rework (PR #6430). Use it as a playbook when porting other tables.
## Migration Candidates
Current `mantine-react-table` consumers that could be migrated:
**Widgets (in `packages/widgets/`):**
- `docker/component.tsx` — Docker container table
- `media-server/component.tsx` — Media server sessions
**Admin pages (in `apps/nextjs/src/app/[locale]/manage/`):**
- `tools/docker/docker-table.tsx`
- `tools/tasks/_components/tasks-table.tsx`
- `tools/api/components/api-keys.tsx`
- `users/_components/user-list.tsx`
- `users/invites/_components/invite-list.tsx`
- `tools/kubernetes/` — 8 Kubernetes resource tables
**Modals:**
- `packages/modals-collection/src/search-engines/request-media-modal.tsx`
**Shared hook to remove once all migrated:**
- `packages/ui/src/hooks/use-translated-mantine-react-table.ts`
---
## 1. Core Migration: mantine-react-table → mantine-datatable
### Import changes
```typescript
// BEFORE
import { MantineReactTable, useMantineReactTable } from "mantine-react-table";
import type { MRT_ColumnDef } from "mantine-react-table";
// AFTER
import { DataTable, useDataTableColumns } from "mantine-datatable";
import type { DataTableColumn, DataTableSortStatus } from "mantine-datatable";
```
### Column definition shape
```typescript
// BEFORE (MRT)
const columns: MRT_ColumnDef<MyRow>[] = [
{
accessorKey: "name",
header: "Name",
size: 200,
enableSorting: true,
Cell: ({ row }) => <Text>{row.original.name}</Text>,
},
];
// AFTER (mantine-datatable)
const columns: DataTableColumn<MyRow>[] = [
{
accessor: "name", // not accessorKey
title: "Name", // not header
width: 200, // not size
sortable: true, // not enableSorting
render: (record) => <Text>{record.name}</Text>, // not Cell
},
];
```
Key differences:
- `accessorKey` → `accessor`
- `header` → `title`
- `size` → `width`
- `Cell: ({ row }) =>` → `render: (record) =>`
- `enableSorting` → `sortable`
- `enableColumnActions` → does not exist (no column action menus)
- `mantineTableBodyCellProps` → `cellsStyle` in `defaultColumnProps`
### Table component
```typescript
// BEFORE
const table = useTranslatedMantineReactTable({ columns, data, ... });
return <MantineReactTable table={table} />;
// AFTER
return (
<DataTable
records={data}
columns={columns}
sortStatus={sortStatus}
onSortStatusChange={setSortStatus}
...
/>
);
```
### Sorting is manual
mantine-datatable does NOT sort for you. You must sort the data yourself:
```typescript
const [sortStatus, setSortStatus] = useState<DataTableSortStatus<MyRow>>({
columnAccessor: "name",
direction: "asc",
});
const sortedData = useMemo(() => {
const mult = sortStatus.direction === "desc" ? -1 : 1;
return [...data].toSorted((a, b) => {
const aVal = a[sortStatus.columnAccessor];
const bVal = b[sortStatus.columnAccessor];
if (typeof aVal === "string") return aVal.localeCompare(bVal as string) * mult;
return ((aVal as number) - (bVal as number)) * mult;
});
}, [data, sortStatus]);
// Pass sortedData to records, not data
<DataTable records={sortedData} sortStatus={sortStatus} onSortStatusChange={setSortStatus} />
```
---
## 2. Making Tables Transparent (Widget Context)
Widgets sit inside board cards that have user-controlled opacity. The table must be fully transparent so the card background shows through. mantine-datatable defaults to `var(--mantine-color-body)` which is opaque.
Create a CSS file alongside the component:
```css
/* styles.css */
.my-table,
.my-table .mantine-datatable-table,
.my-table .mantine-datatable-table thead,
.my-table .mantine-datatable-table tbody,
.my-table .mantine-datatable-table tfoot,
.my-table th,
.my-table td {
background-color: transparent !important;
}
.my-table .mantine-datatable-table tr {
background-color: transparent;
}
/* Hover: translucent instead of opaque */
.my-table .mantine-datatable-table tbody tr:hover {
background-color: color-mix(in srgb, var(--mantine-color-default-hover) 40%, transparent) !important;
}
/* Headers: frosted glass effect */
.my-table th {
white-space: nowrap;
background-color: color-mix(in srgb, var(--mantine-color-body) 60%, transparent) !important;
backdrop-filter: blur(8px);
}
/* Cell overflow */
.my-table td {
white-space: nowrap;
overflow: hidden;
text-overflow: ellipsis;
}
/* Hide sort icon until hover */
.my-table .mantine-datatable-header-cell-sortable-icon {
opacity: 0;
transition: opacity 150ms;
}
.my-table th:hover .mantine-datatable-header-cell-sortable-icon {
opacity: 1;
}
/* Row expansion area */
.my-table .mantine-datatable-row-expansion-cell {
background-color: color-mix(in srgb, var(--mantine-color-body) 40%, transparent) !important;
}
```
Import it: `import "./styles.css";` and apply via `className="my-table"`.
---
## 3. Responsive Column Visibility
Use a lookup table of width breakpoints instead of media queries. The widget receives `width` as a prop.
```typescript
interface SizeConfig {
fontSize: "xs" | "sm";
iconSize: number;
cellPadding: number;
showSpeedColumns: boolean;
showTimeColumn: boolean;
showStateColumn: boolean;
}
const SIZE_BREAKPOINTS: { maxWidth: number; config: SizeConfig }[] = [
{ maxWidth: 300, config: { fontSize: "xs", iconSize: 12, cellPadding: 2, showSpeedColumns: false, showTimeColumn: false, showStateColumn: false } },
{ maxWidth: 500, config: { fontSize: "xs", iconSize: 14, cellPadding: 4, showSpeedColumns: false, showTimeColumn: true, showStateColumn: false } },
];
const DEFAULT_SIZE_CONFIG: SizeConfig = { fontSize: "sm", iconSize: 14, cellPadding: 4, showSpeedColumns: true, showTimeColumn: true, showStateColumn: true };
function getSizeConfig(width: number): SizeConfig {
for (const { maxWidth, config } of SIZE_BREAKPOINTS) {
if (width < maxWidth) return config;
}
return DEFAULT_SIZE_CONFIG;
}
// Usage: memoize it
const size = useMemo(() => getSizeConfig(width), [width]);
```
Then use a visibility check lookup table for columns:
```typescript
const columnVisibilityChecks: Partial<Record<string, (ctx: ColumnContext) => boolean>> = {
upSpeed: (ctx) => ctx.hasTorrents && ctx.size.showSpeedColumns,
downSpeed: (ctx) => ctx.size.showSpeedColumns,
time: (ctx) => ctx.size.showTimeColumn,
state: (ctx) => ctx.size.showStateColumn,
};
function shouldIncludeColumn(accessor: string, ctx: ColumnContext): boolean {
if (!ctx.optionsColumnSet.has(accessor)) return false;
const check = columnVisibilityChecks[accessor];
if (check && !check(ctx)) return false;
return true;
}
```
---
## 4. Column Persistence (Drag/Resize → Server)
`useDataTableColumns` stores column order and widths in localStorage by default. For widgets, this must persist server-side so it works across devices.
### Hidden widget options
In `index.ts`, add hidden text options:
```typescript
columnOrder: factory.text({ defaultValue: "" }),
columnWidths: factory.text({ defaultValue: "" }),
```
With visibility override:
```typescript
{
columnOrder: { shouldHide: () => true },
columnWidths: { shouldHide: () => true },
}
```
### Translation keys required
Even hidden options need translation keys or the translation test fails:
```json
"widget.mywidget.option.columnOrder.label": "Column order",
"widget.mywidget.option.columnWidths.label": "Column widths"
```
### Old-import mappings
Add to `packages/old-import/src/widgets/options.ts`:
```typescript
columnOrder: () => undefined,
columnWidths: () => undefined,
```
### Dynamic storeKey
Use a key that includes `itemId` and sorted column names. This prevents cross-widget localStorage bleeding:
```typescript
const storeKey = `mytable-${itemId ?? "preview"}-${[...options.columns].toSorted().join(",")}`;
```
### Hydration from server → localStorage
On mount, seed `useDataTableColumns` with server-persisted values. Use a `hydrated` ref to prevent the persist effect from immediately writing defaults back:
```typescript
const { effectiveColumns, columnsOrder, columnsWidth, setColumnsOrder, setMultipleColumnWidths } =
useDataTableColumns<MyRow>({ key: storeKey, columns });
const lastStoreKey = useRef(storeKey);
const hydrated = useRef(false);
useEffect(() => {
if (lastStoreKey.current !== storeKey) {
lastStoreKey.current = storeKey;
hydrated.current = false;
}
if (hydrated.current) return;
if (savedOrder.length > 0) setColumnsOrder(savedOrder);
if (Object.keys(savedWidths).length > 0) {
setMultipleColumnWidths(
Object.entries(savedWidths).map(([accessor, w]) => ({ accessor, width: w })),
);
}
requestAnimationFrame(() => { hydrated.current = true; });
}, [storeKey, savedOrder, savedWidths, setColumnsOrder, setMultipleColumnWidths]);
```
### Persist localStorage → server
Watch for changes and write back:
```typescript
const prevOrder = useRef(columnsOrder);
const prevWidths = useRef(columnsWidth);
useEffect(() => {
if (!hydrated.current) return;
const orderChanged = JSON.stringify(columnsOrder) !== JSON.stringify(prevOrder.current);
const widthsChanged = JSON.stringify(columnsWidth) !== JSON.stringify(prevWidths.current);
prevOrder.current = columnsOrder;
prevWidths.current = columnsWidth;
if (!orderChanged && !widthsChanged) return;
if (orderChanged) persistOption({ columnOrder: JSON.stringify(columnsOrder) });
if (widthsChanged) {
const widthMap: Record<string, number> = {};
for (const entry of columnsWidth) {
const key = Object.keys(entry)[0];
if (!key) continue;
const w = entry[key as keyof typeof entry];
if (typeof w === "number") widthMap[key] = w;
else if (typeof w === "string" && w.endsWith("px")) widthMap[key] = parseInt(w, 10);
}
persistOption({ columnWidths: JSON.stringify(widthMap) });
}
}, [columnsOrder, columnsWidth, persistOption]);
```
The `persistOption` helper calls both `setOptions` (local state) and `saveItemOptions` (server mutation).
---
## 5. Context Menu (Right-Click)
Do NOT install `mantine-contextmenu`. Use Mantine's `Menu` + `Portal` directly:
```typescript
interface ContextMenuState {
x: number;
y: number;
item: MyRow;
}
const [contextMenu, setContextMenu] = useState<ContextMenuState | null>(null);
const handleContextMenu = useCallback(
({ record, event }: { record: MyRow; event: React.MouseEvent }) => {
event.preventDefault();
setContextMenu({ x: event.clientX, y: event.clientY, item: record });
},
[],
);
// On DataTable:
<DataTable onRowContextMenu={handleContextMenu} />
// Render:
{contextMenu && <RowContextMenu state={contextMenu} onClose={closeContextMenu} />}
```
The menu component uses `Portal` for correct z-index:
```tsx
function RowContextMenu({ state, onClose }: { state: ContextMenuState; onClose: () => void }) {
return (
<Portal>
<Menu opened onClose={onClose} closeOnItemClick={false} position="right-start" offset={0}>
<Menu.Target>
<Box style={{ position: "fixed", left: state.x, top: state.y, width: 0, height: 0 }} />
</Menu.Target>
<Menu.Dropdown>
<Menu.Item onClick={() => { /* action */ onClose(); }}>Action</Menu.Item>
</Menu.Dropdown>
</Menu>
</Portal>
);
}
```
**Key fix:** Use `closeOnItemClick={false}` if any menu item needs a two-step confirmation (like delete → confirm). Close the menu on scroll: `onScroll={() => { if (contextMenu) closeContextMenu(); }}`.
---
## 6. Row Expansion
```typescript
<DataTable
rowExpansion={{
trigger: "click",
allowMultiple: true,
collapseProps: {
transitionDuration: 200,
animateOpacity: true,
transitionTimingFunction: "ease-out",
},
content: ({ record, collapse }) => <ExpandedRow item={record} collapse={collapse} />,
}}
/>
```
**Critical:** Do NOT set `onRowClick` when using `rowExpansion` with `trigger: "click"`. They conflict and the expansion won't work.
---
## 7. Common Pitfalls & Fixes
### Status indicator dots
Don't use `<Badge variant="dot" />` without children — it renders a pill with empty space. Use a plain `Box`:
```tsx
<Box
w={8}
h={8}
style={{
borderRadius: "50%",
backgroundColor: `var(--mantine-color-${color}-filled)`,
flexShrink: 0,
}}
/>
```
### Tooltip overflow on long text columns
When a column has `ellipsis: true`, the tooltip target element can still extend beyond the cell bounds if inner elements don't clip. Fix:
```tsx
<Tooltip position="bottom-start" ...>
<Group style={{ overflow: "hidden" }}>
...
</Group>
</Tooltip>
```
Use `position="bottom-start"` instead of `position="right"` so the tooltip never escapes the widget horizontally.
### Dark mode tooltips
Mantine tooltips default to the theme's tooltip color which may be white in light mode and stay white in dark mode. Always set `color="dark"` on informational tooltips inside widgets.
### Division by zero
Guard all ratio calculations:
```typescript
if (item.size > 0) ratio = item.sent / item.size;
if (totalDown > 0) ratio = totalUp / totalDown;
```
### Composite idAccessor
When rows can come from multiple integrations, IDs may collide. Use a composite key:
```typescript
idAccessor={(record) => `${record.integration.id}:${record.id}`}
```
### JSON option parsing
Always wrap in try/catch:
```typescript
function parseJsonOption<T>(value: string | undefined, fallback: T): T {
if (!value) return fallback;
try { return JSON.parse(value) as T; }
catch { return fallback; }
}
```
### Clipboard API
`navigator.clipboard.writeText()` can throw in non-HTTPS contexts. Always catch:
```typescript
void navigator.clipboard.writeText(text).catch(() => {});
```
### Sync sortStatus with widget options
If the widget has configurable default sort, sync it:
```typescript
useEffect(() => {
setSortStatus({ columnAccessor: options.defaultSort, direction: defaultSortDirection });
}, [options.defaultSort, defaultSortDirection]);
```
---
## 8. Accessibility
- All `ActionIcon` components need `aria-label`
- Filter toggles (client badges, status badges) need `aria-pressed={isActive}`
- `textSelectionDisabled` on DataTable prevents accidental text selection during drag/resize
---
## 9. Error Handling
- Every mutation needs an `onError` callback with `showErrorNotification`
- The `saveItemOptions` mutation (column persistence) also needs `onError`
- Check `isError` on the main query and show a fallback UI
- Use a shared `mutationOptions` object to avoid repeating `onSettled`/`onError`:
```typescript
const mutationOptions = {
onSettled: () => void utils.myQuery.invalidate(),
onError: () => showErrorNotification({ title: t("errors.actionFailed"), message: t("errors.actionFailedMessage") }),
};
```
---
## 10. i18n
- No hardcoded strings. Every user-visible string goes through `useScopedI18n`
- Column titles: `t("items.{accessor}.columnTitle")`
- Detail labels: `t("items.{accessor}.detailsTitle")`
- States: `t("states.{state}")`
- Actions: `t("actions.{action}")`
- Errors: `t("errors.{errorKey}")`
- Hidden options still need `.label` translation keys
---
## 11. Edit Mode
Disable pointer events on the table during edit mode to prevent conflicts with the board's drag-and-drop:
```typescript
let tablePointerEvents: React.CSSProperties["pointerEvents"];
if (isEditMode) tablePointerEvents = "none";
let rowContextMenuHandler = handleContextMenu;
if (isEditMode) rowContextMenuHandler = undefined;
<DataTable style={{ pointerEvents: tablePointerEvents }} />
```
---
## 12. DataTable Props Reference (Commonly Used)
```typescript
<DataTable
records={sortedData}
columns={effectiveColumns}
storeColumnsKey={storeKey}
withTableBorder={false}
borderRadius={0}
highlightOnHover
striped="odd"
stripedColor={{ dark: "dark.7", light: "gray.0" }}
highlightOnHoverColor={{ dark: "dark.5", light: "gray.1" }}
verticalAlign="center"
fetching={isFetching && data.length === 0}
loaderBackgroundBlur={2}
fz={size.fontSize}
textSelectionDisabled
sortStatus={sortStatus}
onSortStatusChange={setSortStatus}
noRecordsText={t("errors.noItems")}
idAccessor={(record) => record.uniqueKey}
height="100%"
className="my-table"
defaultColumnProps={{
noWrap: true,
draggable: true,
resizable: true,
cellsStyle: () => ({ padding: `${size.cellPadding}px 8px` }),
}}
scrollAreaProps={{ type: "auto", scrollbarSize: 6 }}
rowBackgroundColor={(record) => conditionalBgMap[record.state]}
onRowContextMenu={handleContextMenu}
rowExpansion={{ ... }}
onScroll={() => { if (contextMenu) closeContextMenu(); }}
/>
```
---
## 13. Layout Stability
- Don't use `table-layout: fixed` — it fights mantine-datatable's resize logic
- Give non-name columns explicit `width` values so only the name column flexes
- Use `ellipsis: true` on the name column — it handles `text-overflow: ellipsis` internally
- Names should truncate at the cell boundary, not a character count, so the layout is stable regardless of content length
- `flexShrink: 0` on badges/icons prevents them from being squeezed by long names
---
## 14. Checklist for Each Migration
1. Replace imports (`mantine-react-table` → `mantine-datatable`)
2. Convert column definitions (`accessorKey` → `accessor`, `header` → `title`, `Cell` → `render`, etc.)
3. Implement manual sorting with `useState` + `useMemo`
4. Add CSS file for transparent backgrounds (if widget context)
5. Add responsive column visibility (if widget context)
6. Add hidden options + hydration logic (if column persistence needed)
7. Add old-import mappings for new options
8. Add translation keys (including hidden option labels)
9. Add context menu if row actions exist
10. Add row expansion if detail view exists
11. Add error handling on all mutations
12. Add accessibility attributes
13. Test at multiple widget widths
14. Update docs in `apps/docs/docs/widgets/` if applicable
15. Run `pnpm turbo typecheck` — catches old-import mismatches
16. Run translation spec — catches missing keys
+103
View File
@@ -0,0 +1,103 @@
---
description: How to expose tRPC procedures as MCP tools for AI clients
globs: packages/api/src/router/**/*.ts
alwaysApply: false
---
# MCP Integration Guide
Every new tRPC procedure that provides useful functionality should be exposed as an MCP tool so AI assistants (Claude, Cursor, etc.) can use it. This is a first-class feature of Homarr.
## Adding MCP to a procedure
Add `.meta({ mcp: { enabled: true, description: "..." } })` to the procedure chain:
```typescript
// Read-only tool (query)
myProcedure: protectedProcedure
.meta({ mcp: { enabled: true, description: "Get all items with their status and metadata" } })
.input(z.object({ limit: z.number().default(50) }))
.query(async ({ ctx, input }) => { ... })
// Write tool (mutation)
myAction: protectedProcedure
.meta({ mcp: { enabled: true, description: "Delete an item by ID. Requires interact permission on the integration" } })
.input(z.object({ id: z.string() }))
.mutation(async ({ ctx, input }) => { ... })
// No-input procedure (z.void) — works fine, no .input() needed
getAll: protectedProcedure
.meta({ mcp: { enabled: true, description: "List all resources" } })
.query(async ({ ctx }) => { ... })
```
The `.meta()` call must come **before** `.input()` and `.query()`/`.mutation()` in the chain.
## Writing good descriptions
The description is the **only thing** an AI sees to decide when and how to use the tool. Write it like you're explaining to a colleague who has never used Homarr.
**Must include:**
- What the tool returns or does
- Which integrations/services it works with (if applicable)
- How to get required IDs (e.g., "Use integration_all to get the integrationId")
- Permission requirements (if non-obvious)
**Good:**
```
"Get calendar events for upcoming and recent media releases. Fetches from all connected Sonarr (TV), Radarr (movies), Lidarr (music), and Readarr (books) integrations. Requires integrationIds from integration_all"
```
**Bad:**
```
"Get calendar events"
```
**For tools that return permission fields**, explain what they mean:
```
"List all integrations. Returns permissions.hasUseAccess (read) and permissions.hasInteractAccess (actions) — false means the API key lacks that permission, not an error"
```
## Registering in the MCP router
After adding `.meta()` to procedures, you must register the router in `packages/api/src/mcp.ts`. This file uses **eager imports** (no `lazy()`) because MCP tool extraction requires synchronous access to all procedure definitions.
```typescript
// packages/api/src/mcp.ts
import { myNewRouter } from "./router/my-new-feature";
export const mcpRouter = createTRPCRouter({
// ... existing routers
myFeature: myNewRouter,
});
```
**If your router is nested** (e.g., widget sub-routers), import and register the sub-router directly — do not import a parent that uses `lazy()`.
## What to expose vs. what to skip
**Expose:**
- Data queries (list, search, get by ID)
- Actions users would ask an AI to do (toggle, create, delete, approve)
- Status checks (health, stats, summaries)
**Skip:**
- Subscriptions (WebSocket-only, not supported by MCP)
- Internal procedures (session management, onboarding steps)
- File upload/download procedures
- Procedures that return non-serializable data (streams, blobs)
## Input schema rules
- Use `z.object({...})` for inputs — this maps cleanly to MCP tool parameters
- `z.void()` (no `.input()`) works — the MCP handler patches these to accept empty objects
- Avoid `z.union()` or `z.discriminatedUnion()` at the top level — AI clients struggle with these
- Use `.default()` on optional fields so the AI doesn't need to guess values
## Checklist for new features
1. Add `.meta({ mcp: { enabled: true, description: "..." } })` to relevant procedures
2. Import and register the router in `packages/api/src/mcp.ts` (eager, no lazy)
3. Write descriptions that explain what the tool does, what it needs, and what it returns
4. If the tool requires an ID from another tool, say which one in the description
5. Test with `curl` against `/api/mcp/mcp` to verify the tool appears in `tools/list`
+8
View File
@@ -0,0 +1,8 @@
version = 1
[[analyzers]]
name = "javascript"
[analyzers.meta]
plugins = ["react"]
environment = ["nodejs"]
+12 -34
View File
@@ -1,24 +1,22 @@
# Since .env is gitignored, you can use .env.example to build a new `.env` file when you clone the repo.
# Keep this file up-to-date when you add new variables to `.env`.
# Keep this file up-to-date when you add new variables to \`.env\`.
# This file will be committed to version control, so make sure not to have any secrets in it.
# If you are cloning this repo, create a copy of this file named `.env` and populate it with your secrets.
# Generate two different 64-character hex values with `openssl rand -hex 32` before starting Homarr.
# Never reuse these values or commit the populated `.env` file.
AUTH_SECRET=
# The below secret is not used anywhere but required for Auth.js (Would encrypt JWTs and Mail hashes, both not used)
AUTH_SECRET="supersecret"
# The below secret is used to encrypt integration secrets in the database.
# Homarr will reject an empty or invalid value and print a randomly generated candidate.
SECRET_ENCRYPTION_KEY=
# It should be a 32-byte string, generated by running `openssl rand -hex 32` on Unix
# or starting the project without any (which will show a randomly generated one).
SECRET_ENCRYPTION_KEY=0000000000000000000000000000000000000000000000000000000000000000
LOG_LEVEL='info'
# This is how you can use the sqlite driver:
DB_DRIVER='better-sqlite3'
# Safe local source-development path. SQLite files are ignored by Git.
# Container images use /appdata/db/db.sqlite by default instead.
DB_URL='./db.sqlite'
DB_URL='FULL_PATH_TO_YOUR_SQLITE_DB_FILE'
# Those are the two ways to use the mysql2 driver:
# 1. Using the URL format:
@@ -57,39 +55,19 @@ TURBO_TELEMETRY_DISABLED=1
# Example: DOCKER_SOCKET_PATHS=/var/run/docker.sock,/run/user/1000/podman/podman.sock
# DOCKER_SOCKET_PATHS=
# Redis DB index for the Next.js cache handler (default: 1).
# Must differ from REDIS_DATABASE_INDEX (default: 0) to avoid cache ops affecting app data.
# REDIS_CACHE_DB=1
# Enable kubernetes tool
# ENABLE_KUBERNETES=true
# Enable mock integration
UNSAFE_ENABLE_MOCK_INTEGRATION=true
# Public URL contract. Values must be absolute HTTP(S) URLs without credentials,
# query strings, or fragments. Trailing slashes are normalized at startup.
# Public Homarr origin used by MCP OAuth metadata and redirects behind proxies.
# BASE_URL=https://homarr.example.com
HOMARR_WEBSITE_URL=https://homarr.dev
# PocketBase API used by Homarr and by the standalone documentation site.
# Defaults to HOMARR_WEBSITE_URL.
# WORKSHOP_API_URL=https://homarr.dev
# Public Workshop website used by links in Homarr, Custom Widgets, onboarding,
# and assistant context. Defaults to HOMARR_WEBSITE_URL/workshop.
# WORKSHOP_WEB_URL=https://homarr.dev/workshop
# Documentation and PocketBase deployment settings.
# WORKSHOP_PUBLIC_ORIGIN=https://homarr.dev
# PB_ALLOWED_ORIGINS=*
# Local Workshop stack:
# HOMARR_WEBSITE_URL=http://127.0.0.1:3003
# WORKSHOP_API_URL=http://127.0.0.1:8090
# WORKSHOP_WEB_URL=http://127.0.0.1:3003/workshop
# WORKSHOP_PUBLIC_ORIGIN=http://127.0.0.1:3003
# Optional: GITHUB_CLIENT_ID=
# Optional: GITHUB_CLIENT_SECRET=
# Enable demo mode (seeds a demo user "demo"/"demo", a sample board and a mock integration)
# DEMO_MODE=true
# Whether demo mode is read-only. Defaults to true (the public demo blocks all mutations).
# Set to false for writable preview deployments so the demo user can modify the dashboard.
# DEMO_READ_ONLY=false
# DEMO_READ_ONLY=false
-4
View File
@@ -33,10 +33,6 @@ body:
options:
# The below comment is used to insert a new version with on-release.yml
#NEXT_VERSION#
- 1.76.2
- 1.76.1
- 1.76.0
- 1.75.0
- 1.74.0
- 1.73.0
- 1.72.0
@@ -20,7 +20,7 @@ runs:
using: "composite"
steps:
- name: Prebuilt debian dependencies
uses: ./.github/actions/prebuilt-debian
uses: homarr-labs/homarr/.github/actions/prebuilt-debian@dev
id: prebuilt-debian
with:
architecture: ${{ inputs.architecture }}
@@ -11,6 +11,9 @@ outputs:
runs:
using: "composite"
steps:
- uses: actions/checkout@v7
with:
persist-credentials: false
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v4
Binary file not shown.

Before

Width:  |  Height:  |  Size: 582 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 649 KiB

+8 -8
View File
@@ -4,18 +4,18 @@
<h3>Homarr</h3>
</div>
**Thank you for your contribution. Please ensure that your pull request meets the following requirements:**
**Thank you for your contribution. Please ensure that your pull request meets the following pull request:**
- [ ] Relevant checks pass without warnings or errors
- [ ] Pull request targets the intended base branch (`dev` by default; `release/v2` for v2 release work)
- [ ] Builds without warnings or errors (`pnpm build`, autofix with `pnpm format:fix`)
- [ ] Pull request targets `dev` branch
- [ ] Commits follow the [conventional commits guideline](https://www.conventionalcommits.org/en/v1.0.0/)
- [ ] No shorthand variable names are used (for example, `x`, `y`, `i`, or abbreviations)
- [ ] User-facing changes include matching documentation updates under `apps/docs`
- [ ] When using AI, no temporary files are checked in and the code style follows the rest of the project
- [ ] No shorthand variable names are used (eg. `x`, `y`, `i` or any abbrevation)
- [ ] Documentation is up to date. Create a pull request [here](https://github.com/homarr-labs/documentation/).
- [ ] When using AI; No temp files are checked in, the code style follows the rest of the project
**When contributing a new integration or widget:**
- [ ] I confirmed manually that the integration or widget is working with a real system and performed basic smoke
tests (e.g. what happens if it is offline)
tests (e.g. what happens if it is offline)
- [ ] Optional but recommended: I confirm that I am willing to be added to the `CODEOWNERS` of this feature and will
provide support and maintain the integration (e.g. in case of breaking or failure)
provide support and maintain the integration (e.g. in case of breaking or failure)
+22 -10
View File
@@ -1,35 +1,47 @@
name: Formatting
name: Auto-fix lint and format
on:
pull_request:
types: [opened, synchronize, reopened]
branches: ["**"]
branches: ["*"]
permissions:
contents: read
contents: write
concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
cancel-in-progress: true
jobs:
format:
name: Check formatting
autofix:
name: Auto-fix
if: github.event.pull_request.head.repo.id == github.repository_id
runs-on: ubuntu-latest
timeout-minutes: 5
steps:
- uses: actions/checkout@v7
with:
fetch-depth: 0
persist-credentials: false
ref: ${{ github.head_ref }}
- name: Setup
uses: ./tooling/github/setup
- name: Check changed files
- name: Get changed files
id: changed
env:
BASE_REF: ${{ github.base_ref }}
run: |
git diff --name-only --diff-filter=ACMR -z "origin/${BASE_REF}...HEAD" -- \
'*.css' '*.js' '*.json' '*.jsonc' '*.jsx' '*.md' '*.mdx' '*.scss' '*.ts' '*.tsx' '*.yaml' '*.yml' \
| xargs -0 -r pnpm exec oxfmt --check --
git diff --name-only --diff-filter=ACMR "origin/${BASE_REF}...HEAD" -- '*.ts' '*.tsx' '*.js' '*.jsx' > /tmp/changed-files
echo "count=$(wc -l < /tmp/changed-files | tr -d '[:space:]')" >> $GITHUB_OUTPUT
- name: Auto-fix lint + format
if: steps.changed.outputs.count != '0'
run: |
xargs -r -a /tmp/changed-files pnpm exec oxlint --fix || true
xargs -r -a /tmp/changed-files pnpm exec oxfmt --write
- name: Commit fixes
uses: stefanzweifel/git-auto-commit-action@v7
with:
commit_message: "style: auto-fix lint and format"
@@ -20,7 +20,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_APPROVE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_APPROVE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_APPROVE_APP_ID }}
permission-pull-requests: write # required to enable auto-merge
- name: Install GitHub CLI
+1 -1
View File
@@ -28,7 +28,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_APPROVE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_APPROVE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_APPROVE_APP_ID }}
permission-pull-requests: write # required to approve a pull request and enable auto-merge
- name: Install GitHub CLI
+73 -216
View File
@@ -3,12 +3,12 @@ name: CI
on:
pull_request:
types: [opened, synchronize, reopened, ready_for_review]
branches: ["**"]
branches: ["*"]
pull_request_target:
types: [closed]
branches: ["**"]
branches: ["*"]
push:
branches: [main, dev, beta, next, v2, release/v2]
branches: [main, dev, beta, next]
merge_group:
workflow_dispatch:
inputs:
@@ -34,10 +34,8 @@ permissions:
contents: read
concurrency:
# Every event for a PR shares one group so a close event cancels stale image
# publication and a newer synchronize run cannot be overwritten by an older run.
group: ${{ github.workflow }}-${{ github.event.pull_request.number || inputs.pr_number || github.ref }}
cancel-in-progress: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' || github.event_name == 'workflow_dispatch' || github.ref != 'refs/heads/main' }}
group: ${{ github.workflow }}-${{ github.event_name }}-${{ github.event.pull_request.number || inputs.pr_number || github.ref }}
cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
env:
FORCE_COLOR: 3
@@ -51,7 +49,7 @@ jobs:
name: Fast gate
if: github.event_name != 'pull_request_target' && (github.event_name != 'workflow_dispatch' || inputs.operation != 'cleanup')
runs-on: ubuntu-latest
timeout-minutes: 20
timeout-minutes: 15
env:
TURBO_API: ${{ secrets.TURBO_API }}
TURBO_TEAM: ${{ secrets.TURBO_TEAM }}
@@ -89,34 +87,21 @@ jobs:
- name: Validate workspace configuration
run: pnpm lint:ws
- name: Validate Custom Widget architecture
run: pnpm check:custom-widgets
- name: Validate checked-in OpenAPI schema
run: pnpm check:openapi
- name: Validate Custom Widget bundle budget
run: |
pnpm turbo run build --filter=@homarr/nextjs
pnpm check:custom-widget-bundle
- name: Run unit tests
run: pnpm test
container:
name: Container build
# Same-repository previews already build amd64; only build a fallback when no digest is available.
needs: [fast-gate, preview-amd64]
name: Container and E2E
needs: fast-gate
if: |
!cancelled() && needs.fast-gate.result == 'success' &&
needs.preview-amd64.result != 'failure' && needs.preview-amd64.outputs.digest == '' &&
(
github.event_name == 'merge_group' ||
(github.event_name == 'push' && github.ref == 'refs/heads/release/v2') ||
(github.event_name == 'pull_request' && github.event.pull_request.draft == false) ||
(github.event_name == 'workflow_dispatch' && (inputs.operation == 'validate' || inputs.operation == 'publish'))
)
github.event_name == 'merge_group' ||
(github.event_name == 'pull_request' && github.event.pull_request.draft == false) ||
(github.event_name == 'workflow_dispatch' && (inputs.operation == 'validate' || inputs.operation == 'publish'))
runs-on: ubuntu-latest
timeout-minutes: 20
timeout-minutes: 30
permissions:
contents: read
packages: write
steps:
- uses: actions/checkout@v7
if: "!env.ACT"
@@ -124,12 +109,15 @@ jobs:
persist-credentials: false
ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || '' }}
- name: Setup
uses: ./tooling/github/setup
- name: Set up Docker Buildx
if: needs.preview-amd64.outputs.digest == '' && !env.ACT
if: "!env.ACT"
uses: docker/setup-buildx-action@v4
- name: Build container once
if: needs.preview-amd64.outputs.digest == '' && !env.ACT
if: "!env.ACT"
uses: docker/build-push-action@v7
with:
context: .
@@ -137,13 +125,15 @@ jobs:
platforms: linux/amd64
load: true
push: false
tags: homarr-ci
tags: |
homarr-ci
homarr-e2e
labels: |
org.opencontainers.image.source=${{ github.server_url }}/${{ github.repository }}
org.opencontainers.image.revision=${{ github.sha }}
org.opencontainers.image.ref.name=pr-${{ github.event.pull_request.number || inputs.pr_number || 'none' }}
cache-from: type=gha,scope=homarr-e2e
cache-to: type=gha,scope=homarr-e2e,mode=max
cache-from: type=gha,scope=homarr-linux-amd64
cache-to: type=gha,scope=homarr-linux-amd64,mode=max
secrets: |
TURBO_API=${{ secrets.TURBO_API }}
TURBO_TEAM=${{ secrets.TURBO_TEAM }}
@@ -165,191 +155,66 @@ jobs:
fi
done
DOCKER_BUILDKIT=1 docker build --platform linux/amd64 \
--tag homarr-ci "${secret_args[@]}" .
--tag homarr-ci --tag homarr-e2e "${secret_args[@]}" .
preview-amd64:
name: Build amd64 preview image
needs: fast-gate
if: |
(
(github.event_name == 'pull_request' && github.event.pull_request.draft == false && github.event.pull_request.head.repo.full_name == github.repository) ||
(github.event_name == 'push' && github.ref == 'refs/heads/release/v2') ||
(github.event_name == 'workflow_dispatch' && inputs.operation == 'publish')
)
runs-on: ubuntu-latest
timeout-minutes: 20
permissions:
contents: read
packages: write
outputs:
digest: ${{ steps.preview-build.outputs.digest }}
steps:
- uses: actions/checkout@v7
- name: Cache Playwright browsers
if: "!env.ACT"
id: playwright-cache
uses: actions/cache@v6
with:
persist-credentials: false
ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || '' }}
path: ~/.cache/ms-playwright
key: playwright-${{ runner.os }}-${{ hashFiles('pnpm-lock.yaml') }}
- name: Log in to GHCR
if: "!env.ACT"
uses: docker/login-action@v4
- name: Install Playwright Chromium
if: steps.playwright-cache.outputs.cache-hit != 'true' || env.ACT
run: pnpm exec playwright install --with-deps chromium
- name: Install Playwright system deps
if: steps.playwright-cache.outputs.cache-hit == 'true'
run: pnpm exec playwright install-deps chromium
- name: Run E2E tests
run: pnpm test:e2e
- name: Upload E2E artifacts
if: failure() && !env.ACT
uses: actions/upload-artifact@v7
with:
registry: ${{ env.REGISTRY }}
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
name: playwright-report
path: e2e/test-results/
retention-days: 7
- name: Set up Docker Buildx
if: "!env.ACT"
uses: docker/setup-buildx-action@v4
- name: Build and push amd64 preview image by digest
id: preview-build
if: "!env.ACT"
uses: docker/build-push-action@v7
with:
context: .
network: host
platforms: linux/amd64
outputs: type=image,"name=${{ env.PREVIEW_IMAGE }}",push-by-digest=true,name-canonical=true,push=true
cache-from: type=gha,scope=homarr-linux-amd64
cache-to: type=gha,scope=homarr-linux-amd64,mode=max
secrets: |
TURBO_API=${{ secrets.TURBO_API }}
TURBO_TEAM=${{ secrets.TURBO_TEAM }}
TURBO_TOKEN=${{ secrets.TURBO_TOKEN }}
TURBO_REMOTE_CACHE_SIGNATURE_KEY=${{ secrets.TURBO_REMOTE_CACHE_SIGNATURE_KEY }}
preview-arm64:
name: Build arm64 preview image
needs: fast-gate
if: |
(
(github.event_name == 'pull_request' && github.event.pull_request.draft == false && github.event.pull_request.head.repo.full_name == github.repository) ||
(github.event_name == 'push' && github.ref == 'refs/heads/release/v2') ||
(github.event_name == 'workflow_dispatch' && inputs.operation == 'publish')
)
runs-on: ubuntu-24.04-arm
timeout-minutes: 20
permissions:
contents: read
packages: write
outputs:
digest: ${{ steps.preview-build.outputs.digest }}
steps:
- uses: actions/checkout@v7
if: "!env.ACT"
with:
persist-credentials: false
ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || '' }}
- name: Log in to GHCR
if: "!env.ACT"
uses: docker/login-action@v4
with:
registry: ${{ env.REGISTRY }}
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Docker Buildx
if: "!env.ACT"
uses: docker/setup-buildx-action@v4
- name: Build and push arm64 preview image by digest
id: preview-build
if: "!env.ACT"
uses: docker/build-push-action@v7
with:
context: .
network: host
platforms: linux/arm64
outputs: type=image,"name=${{ env.PREVIEW_IMAGE }}",push-by-digest=true,name-canonical=true,push=true
cache-from: type=gha,scope=homarr-linux-arm64
cache-to: type=gha,scope=homarr-linux-arm64,mode=max
secrets: |
TURBO_API=${{ secrets.TURBO_API }}
TURBO_TEAM=${{ secrets.TURBO_TEAM }}
TURBO_TOKEN=${{ secrets.TURBO_TOKEN }}
TURBO_REMOTE_CACHE_SIGNATURE_KEY=${{ secrets.TURBO_REMOTE_CACHE_SIGNATURE_KEY }}
publish-preview:
name: Publish multi-platform preview image
needs: [fast-gate, preview-amd64, preview-arm64]
if: |
!cancelled() &&
needs.fast-gate.result == 'success' &&
needs.preview-amd64.result == 'success' &&
needs.preview-arm64.result == 'success' &&
(
(github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository) ||
(github.event_name == 'push' && github.ref == 'refs/heads/release/v2') ||
(github.event_name == 'workflow_dispatch' && inputs.operation == 'publish')
)
runs-on: ubuntu-latest
timeout-minutes: 5
permissions:
contents: read
packages: write
pull-requests: read
env:
PREVIEW_TAG: ${{ github.event_name == 'push' && 'v2' || format('pr-{0}', github.event.pull_request.number || inputs.pr_number) }}
steps:
- name: Log in to GHCR
if: "!env.ACT"
uses: docker/login-action@v4
with:
registry: ${{ env.REGISTRY }}
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Docker Buildx
if: "!env.ACT"
uses: docker/setup-buildx-action@v4
- name: Verify PR head before publication
if: github.event_name == 'pull_request'
uses: actions/github-script@v8
- name: Determine preview tag
id: preview
if: |
!env.ACT &&
(
(github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository) ||
(github.event_name == 'workflow_dispatch' && inputs.operation == 'publish')
)
env:
EXPECTED_HEAD_SHA: ${{ github.event.pull_request.head.sha }}
PR_NUMBER: ${{ github.event.pull_request.number || inputs.pr_number }}
run: echo "tag=pr-${PR_NUMBER}" >> "$GITHUB_OUTPUT"
- name: Log in to GHCR
if: steps.preview.outputs.tag != ''
uses: docker/login-action@v4
with:
script: |
const { data: pull } = await github.rest.pulls.get({
owner: context.repo.owner,
repo: context.repo.repo,
pull_number: context.payload.pull_request.number,
});
if (pull.state !== "open") {
core.setFailed(`PR #${pull.number} is ${pull.state}; refusing to publish its preview.`);
return;
}
if (pull.head.sha !== process.env.EXPECTED_HEAD_SHA) {
core.setFailed(`PR #${pull.number} moved from ${process.env.EXPECTED_HEAD_SHA} to ${pull.head.sha}.`);
}
registry: ${{ env.REGISTRY }}
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Publish preview manifest
if: "!env.ACT"
- name: Push preview image
if: steps.preview.outputs.tag != ''
env:
AMD64_DIGEST: ${{ needs.preview-amd64.outputs.digest }}
ARM64_DIGEST: ${{ needs.preview-arm64.outputs.digest }}
PREVIEW_TAG: ${{ steps.preview.outputs.tag }}
run: |
tags=(-t "${PREVIEW_IMAGE}:${PREVIEW_TAG}")
docker buildx imagetools create "${tags[@]}" \
"${PREVIEW_IMAGE}@${AMD64_DIGEST}" \
"${PREVIEW_IMAGE}@${ARM64_DIGEST}"
docker tag homarr-ci "${PREVIEW_IMAGE}:${PREVIEW_TAG}"
docker push "${PREVIEW_IMAGE}:${PREVIEW_TAG}"
- name: Verify preview manifest platforms
if: "!env.ACT"
run: |
tags=("${PREVIEW_TAG}")
for tag in "${tags[@]}"; do
image="${PREVIEW_IMAGE}:${tag}"
manifest="$(docker buildx imagetools inspect "$image" --raw)"
for architecture in amd64 arm64; do
jq --exit-status --arg architecture "$architecture" \
'any(.manifests[]; .platform.os == "linux" and .platform.architecture == $architecture)' \
<<<"$manifest" >/dev/null
done
docker buildx imagetools inspect "$image"
done
- name: Show Act publication result
if: env.ACT
run: echo "Act validated the publication path; external GHCR writes are disabled locally."
cleanup:
name: Clean up preview image
@@ -386,14 +251,6 @@ jobs:
core.info(`No package version found with tag "${tag}"`);
return;
}
const targetTags = target.metadata?.container?.tags ?? [];
const extraTags = targetTags.filter((targetTag) => targetTag !== tag);
if (extraTags.length > 0) {
core.warning(
`Package version ${target.id} also carries tag(s) ${extraTags.map((targetTag) => `"${targetTag}"`).join(", ")}; refusing cleanup.`,
);
return;
}
await github.rest.packages.deletePackageVersionForOrg({
package_type: "container",
package_name: packageName,
@@ -22,7 +22,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.CROWDIN_APP_PRIVATE_KEY }}
client-id: ${{ vars.CROWDIN_CLIENT_ID }}
client-id: ${{ secrets.CROWDIN_APP_ID }}
permission-contents: write # required to commit to crowdin branch
permission-pull-requests: write # required to create pull request
@@ -52,7 +52,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_APPROVE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_APPROVE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_APPROVE_APP_ID }}
permission-pull-requests: write
- name: Approve PR
env:
@@ -64,7 +64,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_MERGE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_MERGE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_MERGE_APP_ID }}
permission-contents: write # write to beta branch (due to merge)
permission-pull-requests: write # merge pull request
- id: automerge
+12 -64
View File
@@ -13,7 +13,7 @@ on:
type: boolean
required: false
default: true
description: Send notifications when redeploying an existing release
description: Send notifications
permissions:
contents: write # Required to update package.json version
@@ -38,11 +38,10 @@ jobs:
SKIP_RELEASE: ${{ github.event_name == 'workflow_dispatch' || github.ref_name == 'dev' || github.ref_name == 'next' }}
outputs:
version: ${{ steps.read-semver.outputs.version || steps.version-fallback.outputs.version }}
git_ref: ${{ steps.read-git-ref.outputs.ref || steps.read-semver.outputs.version || github.ref }}
git_ref: ${{ steps.read-git-ref.outputs.ref || github.ref }}
skipped: ${{ env.SKIP_RELEASE }}
upstream_sync: ${{ steps.update-upstream.outcome }}
steps:
- run: echo "Manual dispatch redeploys the latest existing release without creating a new one"
- run: echo "Skipping release for workflow_dispatch event"
if: env.SKIP_RELEASE == 'true'
# The below generated version fallback represents a normalized branch name, for example "feature/branch-name" -> "feature-branch-name"
- run: echo "version="$(echo ${{github.ref_name}} | sed 's/[^a-zA-Z0-9\-]/-/g') >> "$GITHUB_OUTPUT"
@@ -55,13 +54,12 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_MERGE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_MERGE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_MERGE_APP_ID }}
permission-contents: write # required to commit package.json & changelog changes, merge them to dev / next and publish the release
- uses: actions/checkout@v7
if: env.SKIP_RELEASE == 'false' || github.ref_name == 'main' || github.ref_name == 'beta'
if: env.SKIP_RELEASE == 'false'
with:
fetch-depth: 0
persist-credentials: false
- uses: pnpm/action-setup@v6
if: env.SKIP_RELEASE == 'false'
@@ -70,29 +68,12 @@ jobs:
with:
node-version: 24.18.0
package-manager-cache: false
- name: Verify major release intent
- run: npm i -g pnpm
if: env.SKIP_RELEASE == 'false'
run: |
latest_tag="$(git describe --tags --abbrev=0)"
latest_major="${latest_tag#v}"
latest_major="${latest_major%%.*}"
package_version="$(jq --raw-output '.version' package.json)"
package_major="${package_version%%.*}"
if (( package_major <= latest_major )); then
exit 0
fi
if git log --format='%s%n%b' "${latest_tag}..HEAD" \
| grep --extended-regexp --quiet '(^[[:alnum:]-]+(\([^)]*\))?!:)|(^BREAKING[ -]CHANGE:)'; then
exit 0
fi
echo "::error::package.json declares ${package_version}, but commits since ${latest_tag} contain no breaking-change signal. Use a conventional ! commit or BREAKING CHANGE footer so Semantic Release creates the intended major version."
exit 1
- name: Install dependencies
if: env.SKIP_RELEASE == 'false'
run: pnpm install --frozen-lockfile
run: |
pnpm install
- name: Run Semantic Release
if: env.SKIP_RELEASE == 'false'
env:
@@ -116,10 +97,8 @@ jobs:
run: |
echo "ref=$(git rev-parse HEAD)" >> "$GITHUB_OUTPUT"
- name: Update dev / next branch
id: update-upstream
if: env.SKIP_RELEASE == 'false'
# Finish publishing the already-created release, then fail in the publish job if synchronization did not work.
continue-on-error: true
continue-on-error: true # Prevent pipeline from failing when merge fails
env:
GITHUB_TOKEN: ${{ steps.obtainToken.outputs.token }}
UPSTREAM_BRANCH: ${{ github.ref_name == 'main' && 'dev' || 'next' }}
@@ -226,14 +205,9 @@ jobs:
runs-on: ubuntu-latest
timeout-minutes: 5
steps:
- uses: actions/checkout@v7
if: needs.release.outputs.skipped == 'false'
with:
ref: ${{ needs.release.outputs.git_ref }}
persist-credentials: false
- name: Extract amd64
if: needs.release.outputs.skipped == 'false'
uses: ./.github/actions/extract-build-artifact
uses: homarr-labs/homarr/.github/actions/extract-build-artifact@dev
with:
digest: "${{ env.GHCR_REPO }}@${{ needs.build-amd64.outputs.digest }}"
architecture: amd64
@@ -247,14 +221,9 @@ jobs:
runs-on: ubuntu-24.04-arm
timeout-minutes: 5
steps:
- uses: actions/checkout@v7
if: needs.release.outputs.skipped == 'false'
with:
ref: ${{ needs.release.outputs.git_ref }}
persist-credentials: false
- name: Extract arm64
if: needs.release.outputs.skipped == 'false'
uses: ./.github/actions/extract-build-artifact
uses: homarr-labs/homarr/.github/actions/extract-build-artifact@dev
with:
digest: "${{ env.GHCR_REPO }}@${{ needs.build-arm64.outputs.digest }}"
architecture: arm64
@@ -277,7 +246,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_MERGE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_MERGE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_MERGE_APP_ID }}
permission-contents: write # required to publish the release
- name: Log in to the Container registry
@@ -314,30 +283,9 @@ jobs:
run: |
gh release edit --repo ${{ github.repository }} ${{ needs.release.outputs.version }} --draft=false
- name: Verify dev / next synchronization
if: needs.release.outputs.skipped == 'false' && needs.release.outputs.upstream_sync != 'success'
run: |
echo "::error::The released commit was not synchronized back to the upstream development branch."
exit 1
- name: Discord notification
if: github.event_name != 'workflow_dispatch' || inputs['send-notifications']
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
with:
args: "Successfully deployed images for branch **${{ github.ref_name }}**. Tagged as **${{env.NEXT_VERSION}}**."
notify-failure:
name: Notify deployment failure
if: ${{ always() && (failure() || cancelled()) && (github.event_name != 'workflow_dispatch' || inputs['send-notifications']) }}
needs: [release, build-amd64, build-arm64, extract-asset-amd64, extract-asset-arm64, publish]
runs-on: ubuntu-latest
timeout-minutes: 2
steps:
- name: Discord failure notification
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
with:
args: "The deployment for branch **${{ github.ref_name }}** failed or was cancelled: [run ${{ github.run_number }}](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }})"
+1 -5
View File
@@ -37,16 +37,12 @@ jobs:
run: pnpm turbo build --filter=@homarr/docs
env:
CI: "true"
- name: Validate exported links and search
run: |
pnpm --filter @homarr/docs validate:links
pnpm --filter @homarr/docs verify:search
- name: Setup Pages
uses: actions/configure-pages@v6
- name: Upload pages artifact
uses: actions/upload-pages-artifact@v5
with:
path: "apps/docs/out"
path: "apps/docs/build"
deploy:
runs-on: ubuntu-latest
+11 -31
View File
@@ -19,10 +19,10 @@ permissions:
jobs:
create-and-merge-pr:
runs-on: ubuntu-latest
timeout-minutes: 10
timeout-minutes: 2
steps:
- name: Discord notification
if: ${{ github.event_name == 'schedule' || inputs.send-notifications }}
if: ${{ github.event.inputs.send-notifications }}
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
@@ -38,30 +38,13 @@ jobs:
with:
token: ${{ github.token }}
branch: dev
noNewCommitBehavior: silent
noVersionBumpBehavior: silent
- name: No release needed
if: ${{ steps.semver.outputs.bump == '' || steps.semver.outputs.bump == 'none' }}
run: echo "No release-worthy commits were found since ${{ steps.semver.outputs.current }}." >> "$GITHUB_STEP_SUMMARY"
- name: Create pull request
id: create-pull-request
if: ${{ steps.semver.outputs.bump != '' && steps.semver.outputs.bump != 'none' }}
run: |
pr_url="$(gh pr list --base main --head dev --state open --json url --jq '.[0].url // empty')"
if [[ -z "$pr_url" ]]; then
pr_url="$(gh pr create \
--title "chore(release): automatic release ${{ steps.semver.outputs.next }}" \
--body "**This is an automatic release**.<br/>Manual action may be required for major bumps.<br/>Detected change to be ``${{ steps.semver.outputs.bump }}``<br/>Bump version from ``${{ steps.semver.outputs.current }}`` to ``${{ steps.semver.outputs.next }}``" \
--base main \
--head dev \
--label automerge)"
fi
echo "url=$pr_url" >> "$GITHUB_OUTPUT"
echo "pr_number=${pr_url##*/}" >> "$GITHUB_OUTPUT"
run: 'gh pr create --title "chore(release): automatic release ${{ steps.semver.outputs.next }}" --body "**This is an automatic release**.<br/>Manual action may be required for major bumps.<br/>Detected change to be ``${{ steps.semver.outputs.bump }}``<br/>Bump version from ``${{ steps.semver.outputs.current }}`` to ``${{ steps.semver.outputs.next }}``" --base main --head dev --label automerge'
env:
GITHUB_TOKEN: "${{ secrets.GITHUB_TOKEN }}"
- name: Discord notification
if: ${{ steps.create-pull-request.outcome == 'success' && (github.event_name == 'schedule' || inputs.send-notifications) }}
if: ${{ github.event.inputs.send-notifications }}
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
@@ -69,29 +52,26 @@ jobs:
args: "Created a release PR ${{ steps.create-pull-request.outputs.url }} for version ${{ steps.semver.outputs.next }} (new behaviour: ${{ steps.semver.outputs.bump }})"
- name: Obtain token
id: obtainApprovalToken
if: ${{ steps.create-pull-request.outcome == 'success' }}
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_APPROVE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_APPROVE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_APPROVE_APP_ID }}
permission-pull-requests: write
- name: Approve PR
if: ${{ steps.create-pull-request.outcome == 'success' }}
env:
GITHUB_TOKEN: ${{ steps.obtainApprovalToken.outputs.token }}
run: |
gh pr review "${{ steps.create-pull-request.outputs.url }}" --approve --body "Automatically approved by GitHub Action"
gh pr review --approve --body "Automatically approved by GitHub Action"
- name: Obtain token
id: obtainMergeToken
if: ${{ steps.create-pull-request.outcome == 'success' }}
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_MERGE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_MERGE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_MERGE_APP_ID }}
permission-contents: write # write to main branch (due to merge)
permission-pull-requests: write # merge pull request
- id: automerge
if: ${{ steps.create-pull-request.outcome == 'success' && steps.semver.outputs.bump != 'major' }}
if: ${{ steps.semver.outputs.bump != 'major' }}
name: automerge
uses: "pascalgn/automerge-action@v0.16.4"
env:
@@ -105,21 +85,21 @@ jobs:
MERGE_REQUIRED_APPROVALS: 0 # do not require approvals
- name: Merged Discord notification
if: ${{ steps.automerge.outputs.mergeResult == 'merged' && (github.event_name == 'schedule' || inputs.send-notifications) }}
if: ${{ steps.automerge.outputs.mergeResult == 'merged' && github.event.inputs.send-notifications }}
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
with:
args: "Merged PR ${{ steps.create-pull-request.outputs.url }} for release ${{ steps.semver.outputs.next }}"
- name: Major Bump Discord notification
if: ${{ steps.semver.outputs.bump == 'major' && (github.event_name == 'schedule' || inputs.send-notifications) }}
if: ${{ steps.semver.outputs.bump == 'major' && github.event.inputs.send-notifications }}
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
with:
args: "The release PR must be manually merged because the next version is a major version: ${{ steps.create-pull-request.outputs.url }} for release ${{ steps.semver.outputs.next }}"
- name: Discord Fail Notification
if: ${{ failure() && (github.event_name == 'schedule' || inputs.send-notifications) }}
if: failure() && github.event.inputs.send-notifications
env:
DISCORD_WEBHOOK: ${{ secrets.DISCORD_WEBHOOK }}
uses: Ilshidur/action-discord@0.4.0
@@ -0,0 +1,58 @@
name: "[Docs] Algolia Recrawl"
on:
workflow_run:
workflows: ["[Deployment] Documentation"]
types: [completed]
branches: [dev]
workflow_dispatch:
permissions:
contents: read
jobs:
recrawl:
if: >-
${{
(github.event.workflow_run.conclusion == 'success' && github.event.workflow_run.event != 'pull_request')
|| github.event_name == 'workflow_dispatch'
}}
runs-on: ubuntu-latest
timeout-minutes: 10
steps:
- uses: actions/checkout@v7
- name: Setup
uses: ./tooling/github/setup
- name: Check Algolia crawler secrets
run: |
missing=0
for secret in ALGOLIA_CRAWLER_ID ALGOLIA_CRAWLER_USER_ID ALGOLIA_CRAWLER_API_KEY; do
if [ -z "${!secret}" ]; then
echo "::warning::Missing repository secret: $secret"
missing=1
fi
done
if [ "$missing" -eq 1 ]; then
echo "Add crawler secrets to enable automatic recrawls."
exit 1
fi
env:
ALGOLIA_CRAWLER_ID: ${{ secrets.ALGOLIA_CRAWLER_ID }}
ALGOLIA_CRAWLER_USER_ID: ${{ secrets.ALGOLIA_CRAWLER_USER_ID }}
ALGOLIA_CRAWLER_API_KEY: ${{ secrets.ALGOLIA_CRAWLER_API_KEY }}
- name: Wait for GitHub Pages to serve the new deployment
run: sleep 60
- name: Trigger Algolia DocSearch recrawl
uses: cssnr/algolia-crawler-action@v2
with:
crawler_id: ${{ secrets.ALGOLIA_CRAWLER_ID }}
crawler_user_id: ${{ secrets.ALGOLIA_CRAWLER_USER_ID }}
crawler_api_key: ${{ secrets.ALGOLIA_CRAWLER_API_KEY }}
- name: Verify search index
working-directory: apps/docs
run: pnpm verify:search
@@ -33,7 +33,7 @@ jobs:
- name: Commit updated contributors
uses: EndBug/add-and-commit@v10
with:
add: "apps/docs/public"
add: "apps/docs/static"
default_author: github_actor
message: "chore(docs): update contributors"
push: true
+31 -3
View File
@@ -26,13 +26,41 @@ jobs:
- name: Copy env
run: cp .env.example .env
- name: Generate and validate OpenAPI spec
run: pnpm scripts:update-openapi
- name: Start dev server
run: pnpm --filter @homarr/nextjs dev &
env:
TURBO_TELEMETRY_DISABLED: 1
DISABLE_TELEMETRY: 1
- name: Wait for server and fetch spec
run: |
set -euo pipefail
for i in $(seq 1 30); do
if curl -fsS http://localhost:3000/api/openapi -o /tmp/openapi.json 2>/dev/null; then
echo "Server is up, spec fetched"
break
fi
echo "Waiting for server... ($i/30)"
sleep 3
done
if [ ! -f /tmp/openapi.json ]; then
echo "::error::Server failed to start within timeout"
exit 1
fi
jq -e '.openapi and .paths' /tmp/openapi.json > /dev/null || {
echo "::error::Fetched file is not a valid OpenAPI spec"
exit 1
}
cp /tmp/openapi.json apps/docs/static/api/open-api-schema.json
echo "OpenAPI spec updated ($(jq '.paths | length' /tmp/openapi.json) paths)"
- name: Commit updated spec
uses: EndBug/add-and-commit@v10
with:
add: "apps/docs/public/api/"
add: "apps/docs/static/api/"
default_author: github_actor
message: "chore(docs): update OpenAPI spec"
push: true
@@ -5,7 +5,7 @@ permissions:
on:
pull_request:
branches: ["**"]
branches: ["*"]
paths: [".github/actions/prebuilt-debian/**", "deployments/prebuilt-debian/**"]
jobs:
@@ -5,7 +5,7 @@ permissions:
on:
pull_request:
branches: ["**"]
branches: ["*"]
paths: [".github/renovate.json5"]
jobs:
+34 -1
View File
@@ -5,6 +5,39 @@ on:
types: [released]
jobs:
trigger-docs-release:
name: Trigger Documentation Release
runs-on: ubuntu-latest
timeout-minutes: 2
steps:
- name: Obtain token
id: obtainToken
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.HOMARR_DOCS_RELEASE_APP_PRIVATE_KEY }}
client-id: ${{ vars.HOMARR_DOCS_RELEASE_APP_ID }}
owner: homarr-labs
repositories: |
documentation
permission-contents: write # required to dispatch repository workflow
- name: Trigger documentation release
env:
GITHUB_TOKEN: ${{ steps.obtainToken.outputs.token }}
SOURCE_TAG: ${{ github.event.release.tag_name }}
run: |
curl -X POST \
-H "Authorization: token $GITHUB_TOKEN" \
-H "Accept: application/vnd.github+json" \
https://api.github.com/repos/homarr-labs/documentation/dispatches \
-d @- <<EOF
{
"event_type": "trigger-release",
"client_payload": {
"tag": "${SOURCE_TAG}"
}
}
EOF
update-bug-report-template:
name: Update Bug Report Template
runs-on: ubuntu-latest
@@ -15,7 +48,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.RENOVATE_MERGE_PRIVATE_KEY }}
client-id: ${{ vars.RENOVATE_MERGE_CLIENT_ID }}
client-id: ${{ secrets.RENOVATE_MERGE_APP_ID }}
permission-contents: write # required to commit to branch
permission-pull-requests: write # required to create pr & enable automerge
- name: Checkout code
@@ -1,6 +1,5 @@
permissions:
statuses: write
contents: read
on:
pull_request:
@@ -17,20 +16,12 @@ jobs:
env:
GH_TOKEN: ${{secrets.GITHUB_TOKEN}}
run: |
for attempt in 1 2 3 4 5; do
if gh api \
--method POST \
-H "Accept: application/vnd.github+json" \
-H "X-GitHub-Api-Version: 2022-11-28" \
/repos/${{github.repository}}/statuses/${{github.event.pull_request.head.sha}} \
-f 'state=success' \
-f 'target_url=https://docs.renovatebot.com/configuration-options#minimumreleaseage' \
-f 'description=Skipped for non renovate branches' \
-f 'context=renovate/stability-days' \
--silent; then
exit 0
fi
echo "status POST failed (attempt $attempt/5), retrying in 5s"
sleep 5
done
exit 1
gh api \
--method POST \
-H "Accept: application/vnd.github+json" \
-H "X-GitHub-Api-Version: 2022-11-28" \
/repos/${{github.repository}}/statuses/${{github.event.pull_request.head.sha}} \
-f 'state=success' \
-f 'target_url=https://docs.renovatebot.com/configuration-options#minimumreleaseage' \
-f 'description=Skipped for non renovate branches' \
-f 'context=renovate/stability-days'
+6 -3
View File
@@ -15,13 +15,16 @@ jobs:
update-contributors:
runs-on: ubuntu-latest
timeout-minutes: 2
strategy:
matrix:
node-version: [22]
steps:
- name: Obtain token
id: obtainToken
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.HOMARR_UPDATE_CONTRIBUTORS_PRIVATE_KEY }}
client-id: ${{ vars.HOMARR_UPDATE_CONTRIBUTORS_CLIENT_ID }}
client-id: ${{ vars.HOMARR_UPDATE_CONTRIBUTORS_APP_ID }}
permission-contents: write # required to commit to branch
permission-pull-requests: write # required to create pr & enable automerge
@@ -31,10 +34,10 @@ jobs:
GITHUB_TOKEN: ${{ steps.obtainToken.outputs.token }}
- uses: pnpm/action-setup@v6
- name: Use Node.js
- name: Use Node.js ${{ matrix.node-version }}
uses: actions/setup-node@v6
with:
node-version: 24.18.0
node-version: ${{ matrix.node-version }}
cache: "pnpm"
- name: Run update script
@@ -24,7 +24,7 @@ jobs:
uses: actions/create-github-app-token@v3
with:
private-key: ${{ secrets.HOMARR_UPDATE_CONTRIBUTORS_PRIVATE_KEY }}
client-id: ${{ vars.HOMARR_UPDATE_CONTRIBUTORS_CLIENT_ID }}
client-id: ${{ vars.HOMARR_UPDATE_CONTRIBUTORS_APP_ID }}
permission-contents: write # required to commit to branch
permission-pull-requests: write # required to create pr & enable automerge
- name: Checkout code
-163
View File
@@ -1,163 +0,0 @@
name: Workshop
on:
pull_request:
paths:
- ".github/workflows/workshop.yml"
- "apps/workshop/**"
- "apps/docs/**"
- "packages/custom-widgets/**"
- "packages/definitions/**"
- "packages/workshop/**"
- "tooling/typescript/**"
- "patches/**"
- "package.json"
- "pnpm-lock.yaml"
- "pnpm-workspace.yaml"
- "turbo.json"
- ".npmrc"
push:
branches: [dev, v2, release/v2, main]
paths:
- ".github/workflows/workshop.yml"
- "apps/workshop/**"
- "apps/docs/**"
- "packages/custom-widgets/**"
- "packages/definitions/**"
- "packages/workshop/**"
- "tooling/typescript/**"
- "patches/**"
- "package.json"
- "pnpm-lock.yaml"
- "pnpm-workspace.yaml"
- "turbo.json"
- ".npmrc"
permissions:
contents: read
concurrency:
group: ${{ github.workflow }}-${{ (github.ref_name == 'v2' || github.ref_name == 'release/v2') && 'v2-preview' || github.ref }}
cancel-in-progress: true
jobs:
validate:
runs-on: ubuntu-24.04
timeout-minutes: 5
steps:
- uses: actions/checkout@v7
- run: docker compose --env-file apps/workshop/.env.staging.example -f apps/workshop/docker-compose.staging.yml config --quiet
publish-amd64:
if: github.event_name == 'push' && (github.ref_name == 'v2' || github.ref_name == 'release/v2' || github.ref_name == 'main')
needs: validate
runs-on: ubuntu-24.04
timeout-minutes: 20
permissions:
contents: read
packages: write
outputs:
digest: ${{ steps.build.outputs.digest }}
steps:
- uses: actions/checkout@v7
- uses: docker/setup-buildx-action@v4
- uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- uses: docker/build-push-action@v7
id: build
with:
context: .
file: apps/workshop/Dockerfile
target: production
platforms: linux/amd64
outputs: type=image,"name=ghcr.io/homarr-labs/workshop",push-by-digest=true,name-canonical=true,push=true
cache-from: type=gha,scope=workshop-linux-amd64
cache-to: type=gha,scope=workshop-linux-amd64,mode=max
publish-arm64:
if: github.event_name == 'push' && (github.ref_name == 'v2' || github.ref_name == 'release/v2' || github.ref_name == 'main')
needs: validate
runs-on: ubuntu-24.04-arm
timeout-minutes: 20
permissions:
contents: read
packages: write
outputs:
digest: ${{ steps.build.outputs.digest }}
steps:
- uses: actions/checkout@v7
- uses: docker/setup-buildx-action@v4
- uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- uses: docker/build-push-action@v7
id: build
with:
context: .
file: apps/workshop/Dockerfile
target: production
platforms: linux/arm64
outputs: type=image,"name=ghcr.io/homarr-labs/workshop",push-by-digest=true,name-canonical=true,push=true
cache-from: type=gha,scope=workshop-linux-arm64
cache-to: type=gha,scope=workshop-linux-arm64,mode=max
publish:
if: github.event_name == 'push' && (github.ref_name == 'v2' || github.ref_name == 'release/v2' || github.ref_name == 'main')
needs: [publish-amd64, publish-arm64]
runs-on: ubuntu-24.04
timeout-minutes: 5
permissions:
contents: read
packages: write
env:
WORKSHOP_IMAGE: ghcr.io/homarr-labs/workshop
steps:
- uses: docker/setup-buildx-action@v4
- uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Verify branch head before publication
uses: actions/github-script@v8
with:
script: |
const branchName = context.ref.replace("refs/heads/", "");
const { data: branch } = await github.rest.repos.getBranch({
owner: context.repo.owner,
repo: context.repo.repo,
branch: branchName,
});
if (branch.commit.sha !== context.sha) {
core.setFailed(`Branch ${branchName} moved from ${context.sha} to ${branch.commit.sha}.`);
}
- name: Publish multi-platform image
env:
AMD64_DIGEST: ${{ needs.publish-amd64.outputs.digest }}
ARM64_DIGEST: ${{ needs.publish-arm64.outputs.digest }}
run: |
tags=(-t "${WORKSHOP_IMAGE}:sha-${GITHUB_SHA}")
if [[ "${GITHUB_REF_NAME}" == "v2" || "${GITHUB_REF_NAME}" == "release/v2" ]]; then
tags+=(-t "${WORKSHOP_IMAGE}:v2")
fi
if [[ "${GITHUB_REF_NAME}" == "main" ]]; then
tags+=(-t "${WORKSHOP_IMAGE}:latest")
fi
docker buildx imagetools create "${tags[@]}" \
"${WORKSHOP_IMAGE}@${AMD64_DIGEST}" \
"${WORKSHOP_IMAGE}@${ARM64_DIGEST}"
- name: Verify multi-platform image
run: |
image="${WORKSHOP_IMAGE}:sha-${GITHUB_SHA}"
manifest="$(docker buildx imagetools inspect "$image" --raw)"
for architecture in amd64 arm64; do
jq --exit-status --arg architecture "$architecture" \
'any(.manifests[]; .platform.os == "linux" and .platform.architecture == $architecture)' \
<<<"$manifest" >/dev/null
done
docker buildx imagetools inspect "$image"
+5 -12
View File
@@ -16,7 +16,6 @@ out/
next-env.d.ts
# artifacts
.ai-evaluations/
packages/db/migrations/*/migrate.cjs
# nitro
@@ -25,8 +24,6 @@ packages/db/migrations/*/migrate.cjs
# production
build
tools/homarr-dev/homarr
tools/homarr-dev/homarr-dev
# misc
.DS_Store
@@ -74,8 +71,7 @@ apps/nextjs/public/images/background.png
# personal agent config
.opencode
.agents/*
!.agents/skills/
.agents/
skills-lock.json
# docusaurus
@@ -83,14 +79,11 @@ apps/docs/.docusaurus
apps/docs/build
.pnpm-store
benchmark-results
# local performance investigation (never ship benchmark data or restored backups)
.widget-performance/
.bench/
# Homarr backups used by scripts/benchmarks/stress-restore.mts. These contain
# real integration secrets and an encryption key — never commit one.
homarr-backup*.zip
*.sqlite
*.sqlite-wal
*.sqlite-shm
/scripts/benchmarks/
# PR description screenshots belong on the pull request, not in the repository
.screenshots/
-8
View File
@@ -1,8 +0,0 @@
{
"mcpServers": {
"mantine": {
"command": "npx",
"args": ["-y", "@mantine/mcp-server@9.6.0"]
}
}
}
+1 -16
View File
@@ -3,20 +3,5 @@
"printWidth": 120,
"endOfLine": "lf",
"sortPackageJson": true,
"ignorePatterns": [
"**/*.cjs",
"**/lang/**",
"static-data/**",
"CHANGELOG.md",
"apps/docs/.docusaurus/**",
"apps/docs/build/**",
"apps/docs/blog/2022/**",
"apps/docs/blog/2023/**",
"apps/docs/blog/2024/**",
"apps/docs/blog/2025/**",
"apps/docs/blog/authors.yml",
"apps/docs/static/api/open-api-schema.json",
"packages/custom-widgets/src/core/component-catalog.generated.json",
"apps/docs/static/custom-widgets/component-catalog-v1.json"
]
"ignorePatterns": ["**/*.cjs", "**/lang/**", "static-data/**", "CHANGELOG.md", "apps/docs/.docusaurus/**", "apps/docs/build/**", "apps/docs/blog/**"]
}
+1 -29
View File
@@ -30,35 +30,7 @@
"options": {
"reportUnusedDisableDirectives": "off"
},
"overrides": [
{
"files": ["apps/nextjs/src/**/*.ts", "apps/nextjs/src/**/*.tsx"],
"excludeFiles": ["**/*.spec.ts", "**/*.spec.tsx"],
"rules": {
"eslint/no-restricted-imports": [
"error",
{
"paths": [
{
"name": "@homarr/widgets",
"allowTypeImports": true,
"message": "Import a narrow @homarr/widgets export to preserve widget lazy loading."
}
]
}
]
}
}
],
"ignorePatterns": [
"**/*.config.js",
"**/*.config.mjs",
"**/*.cjs",
"**/lang/*.d.json.ts",
".next/**",
"apps/docs/.docusaurus/**",
"apps/docs/build/**"
],
"ignorePatterns": ["**/*.config.js", "**/*.config.mjs", "**/*.cjs", "**/lang/*.d.json.ts", ".next/**", "apps/docs/.docusaurus/**", "apps/docs/build/**"],
"settings": {
"react": { "version": "19" }
}
+10 -22
View File
@@ -6,24 +6,23 @@
homarr/
├── apps/
│ ├── nextjs/ # Main Next.js application (port 3000)
│ ├── docs/ # Next.js + Fumadocs documentation site (@homarr/docs)
│ ├── tasks/ # Cron-job initialization and scheduling runtime
│ ├── websocket/ # Standalone tRPC WebSocket server (port 3001)
│ └── workshop/ # Go/PocketBase Workshop service
│ ├── docs/ # Docusaurus 3 documentation site (@homarr/docs)
│ ├── tasks/ # Cron job runner + Fastify tRPC API (port 3002)
│ └── websocket/ # Standalone tRPC WebSocket server (port 3001)
├── packages/
│ ├── api/ # tRPC appRouter, procedures, OpenAPI
│ ├── auth/ # NextAuth config, providers, session, API keys
│ ├── db/ # Drizzle schema (3 DB drivers), migrations, queries
│ ├── core/ # Env validation, DB/Redis driver factories, logging
│ ├── definitions/ # Domain enums: WidgetKind, IntegrationKind, permissions
│ ├── widgets/ # Dashboard widget definitions and components
│ ├── widgets/ # All 39 dashboard widgets (definitions + components)
│ ├── integrations/ # Integration classes (HTTP clients to external apps)
│ ├── redis/ # Redis pub/sub channels, caching abstractions
│ ├── translation/ # next-intl setup, locale configs, lang JSON files
│ ├── ui/ # Shared Mantine components, theme, hooks
│ ├── validation/ # Shared zod schemas for API/forms
│ ├── common/ # Shared utilities, IDs, errors
│ ├── cron-jobs/ # Cron job implementations
│ ├── cron-jobs/ # Cron job implementations (25+ jobs)
│ ├── cron-jobs-core/ # Cron scheduling primitives
│ ├── cron-job-status/ # Cron status via Redis
│ ├── boards/ # Board context, edit mode, cache updater
@@ -41,9 +40,8 @@ homarr/
│ ├── analytics/ # Server-side analytics (Umami)
│ ├── server-settings/ # Server setting keys/types
│ ├── settings/ # User-facing settings UI context
│ ├── custom-widgets/ # Custom JSX v2 schema, validation, and runtime
│ ├── onboarding/ # Onboarding studio and setup flow
│ ├── workshop/ # Homarr-side Workshop client and contracts
│ ├── old-import/ # Legacy Homarr import
│ ├── old-schema/ # Legacy Homarr zod schemas
│ └── cli/ # Node CLI for ops (brocli)
├── tooling/
│ ├── typescript/ # Base tsconfig
@@ -57,13 +55,13 @@ homarr/
## Documentation Sync
The documentation site lives at `apps/docs/` (Next.js + Fumadocs, `@homarr/docs`).
The documentation site lives at `apps/docs/` (Docusaurus 3, `@homarr/docs`).
When modifying user-facing code, you MUST also update the corresponding documentation:
- New integration → `apps/docs/docs/integrations/<slug>/index.mdx` + `index.ts`
- New widget → `apps/docs/docs/widgets/<slug>/index.mdx` + `index.ts`
- Changed API → `apps/docs/docs/management/api/index.mdx`
- Changed API → `apps/docs/docs/management/api.mdx`
- New/changed env vars → `apps/docs/docs/advanced/`
- New CLI commands → `apps/docs/docs/advanced/command-line/`
- Auth changes → `apps/docs/docs/advanced/` SSO pages
@@ -75,7 +73,7 @@ When modifying user-facing code, you MUST also update the corresponding document
- `pnpm dev:cli -- dev` — run the developer CLI without installing a global binary
- `pnpm db:seed` — seed default database data explicitly
- `pnpm docker:dev:up` — start the Redis development service in the background
- `pnpm dev:docs` — Fumadocs site only
- `pnpm dev:docs` — Docusaurus docs site only
- `pnpm turbo build` — build all packages
- `pnpm turbo build --filter=@homarr/docs` — build docs only
- `pnpm turbo typecheck` — typecheck all packages
@@ -86,19 +84,9 @@ When modifying user-facing code, you MUST also update the corresponding document
- Lint: oxlint (not ESLint)
- Format: oxfmt (not Prettier)
- UI: Mantine (not Tailwind) — Tailwind is only used in docs app
- Mantine: use the project-scoped MCP server in `.mcp.json` for current v9 APIs, and check `packages/ui/` before creating a new primitive.
- Icons: @tabler/icons-react
- Docs app can import from `@homarr/definitions` for shared types
- Run `pnpm dev:cli -- dev` to browse local `homarr:*` images and remote PR images.
- Run `pnpm dev:cli -- build <name>` from a Homarr checkout to build `homarr:<name>` with rebuild provenance.
- Run `pnpm dev:cli -- build --pr <number>` to build a PR locally from a temporary checkout.
- Run Go checks from `tools/homarr-dev` with `go test ./...` and `go vet ./...`.
## Agent Skills
Portable skills live in `.agents/skills/`. Read the relevant `SKILL.md` before working in that domain; detailed references are loaded only when needed. Claude-compatible discovery is provided through `.claude/skills`.
- `codebase-context` — architecture, package boundaries, and shared utilities
- `documentation-sync` — required user-facing documentation updates
- `mcp-integration` — safe tRPC-to-MCP exposure
- `homarr-custom-widget` — safe Custom JSX v2 authoring
-8
View File
@@ -1,11 +1,3 @@
## [1.76.2](https://github.com/homarr-labs/homarr/compare/v1.76.1...v1.76.2) (2026-08-27)
## [1.76.1](https://github.com/homarr-labs/homarr/compare/v1.76.0...v1.76.1) (2026-08-26)
## [1.76.0](https://github.com/homarr-labs/homarr/compare/v1.75.0...v1.76.0) (2026-08-21)
## [1.75.0](https://github.com/homarr-labs/homarr/compare/v1.74.0...v1.75.0) (2026-08-14)
## [1.74.0](https://github.com/homarr-labs/homarr/compare/v1.73.0...v1.74.0) (2026-08-07)
## [1.73.0](https://github.com/homarr-labs/homarr/compare/v1.72.0...v1.73.0) (2026-07-31)
+96
View File
@@ -0,0 +1,96 @@
# Changes
## Goal
Remove widget-only services from Homarr's global Integration system.
The affected services are release providers and Search.ch. They are only used by specific widgets, so keeping them as global integrations added unnecessary setup, UI, database rows, docs pages, and integration package code.
ArchiveTeam Warrior remains a proper integration but is no longer auto-seeded on first install.
## Why
These services do not need shared integration state.
- Release providers are selected per Releases widget repository.
- Search.ch is only used by the Timetable widget.
Moving them into widget options makes configuration simpler and removes dead-weight integration code from `packages/integrations`.
## What Changed
### Release Providers
- Added release provider definitions in `packages/definitions/src/release-provider.ts`.
- Removed release providers from `IntegrationKind`.
- Moved release-fetching code out of `packages/integrations` and into `packages/request-handler/src/release-providers.ts`.
- Updated the Releases widget to store provider configuration directly in widget options:
- `provider`
- `identifier`
- `versionFilter`
- `providerUrl`
- Updated the Releases widget editor to use a provider select instead of an integration select.
- Updated Docker import mapping so `ghcr.io` maps to `gitHubContainerRegistry`.
- GitHub Container Registry now uses the public OCI registry API and accepts full image references such as `ghcr.io/homarr-labs/homarr:latest`.
### Timetable / Search.ch
- Removed Search.ch as an integration.
- Added a Timetable widget `baseUrl` option, defaulting to `https://search.ch`.
- Moved Search.ch request/parsing logic into `packages/request-handler/src/timetable.ts`.
- Updated the Timetable API to receive `baseUrl` directly from widget options.
### ArchiveTeam Warrior
- Kept as a proper integration (not removed).
- Removed `defaultUrl` from the integration definition to prevent auto-seeding on first install.
### Integration Cleanup
- Removed creators, exports, docs references, and implementation files for the deleted integrations.
- Removed unused dependencies from `packages/integrations/package.json`.
### Database Migration
- Added `packages/db/migrations/custom/0003_remove_widget_only_integrations.ts`.
- The migration moves old integration-linked widget config into widget options, then deletes the old widget-only integration rows.
- Updated the earlier release widget migration to use the new `provider` field instead of `providerIntegrationId`.
### Docs
- Updated Releases widget docs to describe providers instead of integrations.
- Updated Timetable docs to remove Search.ch integration setup.
- Removed obsolete integration docs pages for the deleted integration kinds.
- Updated docs sitemap and integration docs mappings.
### Translations
- Added English labels for the new widget options:
- `widget.timetable.option.baseUrl.label`
- `widget.releases.option.repositories.providerUrl.label`
## How It Works Now
Release provider configuration lives on each Releases widget repository entry. The request handler resolves provider metadata and fetches releases directly from the provider-specific endpoint. GHCR uses anonymous OCI registry access for public packages.
Timetable passes its widget URL option directly to its API router and request handler. No global integration lookup is needed.
ArchiveTeam Warrior continues to use the standard integration system, but users must manually add the integration (it is not auto-created on install).
## Deliberate Non-Goals
- No provider secret/token support was added.
- No new secret storage model was introduced.
- No compatibility layer was kept in `packages/integrations`.
Widget options are client-visible, so storing provider credentials there would be wrong. If authenticated provider requests are needed later, they need a separate secret-storage design.
## Verification So Far
- Focused typechecks passed for affected packages.
- Full `pnpm turbo typecheck` passed.
- Full `pnpm lint` passed with existing warnings.
- Focused release provider normalization tests passed.
- Widget translation test passed.
- Live GHCR smoke checks passed for `ghcr.io/muchobien/pocketbase:latest` and `ghcr.io/homarr-labs/homarr:latest`.
- `pnpm test packages/integrations/test/pi-hole.spec.ts` passed earlier, but failed on rebased reruns because Pi-hole reported `domainsBeingBlocked` as `-2`/`0` instead of `> 1`.
+21 -2
View File
@@ -22,7 +22,7 @@ COPY --parents ./packages/definitions/src ./
# isolated linker, patches apply inside the virtual store (node_modules/.pnpm/...)
# which IS created by pnpm fetch. The original hoisted linker is restored
# before the install step so the final node_modules layout stays flat.
RUN --mount=type=cache,id=pnpm-store,target=/pnpm/store,sharing=locked \
RUN --mount=type=cache,id=pnpm-store,target=/pnpm/store \
pnpm config set store-dir /pnpm/store && \
sed -i 's/nodeLinker: hoisted/nodeLinker: isolated/' pnpm-workspace.yaml && \
pnpm fetch --ignore-scripts && \
@@ -31,7 +31,7 @@ RUN --mount=type=cache,id=pnpm-store,target=/pnpm/store,sharing=locked \
# Install only from the fetched, committed lockfile so local and Docker builds
# resolve the same dependency graph. Serial lifecycle builds avoid esbuild's
# atomic binary replacement racing across the hoisted workspace (pnpm/pnpm#8200).
RUN --mount=type=cache,id=pnpm-store,target=/pnpm/store,sharing=locked \
RUN --mount=type=cache,id=pnpm-store,target=/pnpm/store \
npm_config_nodedir=/usr/local pnpm install --recursive --offline --frozen-lockfile --child-concurrency=1
COPY . .
@@ -40,6 +40,25 @@ ARG SKIP_ENV_VALIDATION='true'
ARG CI='true'
ARG DISABLE_REDIS_LOGS='true'
ARG TARGETPLATFORM
# Opt-in React profiling build + browser source maps, for React DevTools' Profiler
# and for readable names in heap snapshots. See next.config.ts.
ARG HOMARR_PROFILING='false'
ENV HOMARR_PROFILING=${HOMARR_PROFILING}
# React ships a separate profiling renderer; the production one omits the hooks
# DevTools needs, which is what makes it report "Profiling not supported". Next only
# swaps it via a webpack alias (create-compiler-aliases.js), so under Turbopack both
# `next build --profile` and `reactProductionProfiling` silently produce a normal
# build. Point the vendored client entry at the profiling bundle instead — the same
# swap webpack's alias performs, done where Turbopack cannot miss it.
RUN if [ "$HOMARR_PROFILING" = "true" ]; then \
set -e; \
entry=node_modules/next/dist/compiled/react-dom/client.js; \
grep -q "react-dom-client.production.js" "$entry" || { echo "profiling patch: unexpected $entry"; exit 1; }; \
sed -i "s#./cjs/react-dom-client.production.js#./cjs/react-dom-profiling.profiling.js#" "$entry"; \
grep -q "react-dom-profiling.profiling.js" "$entry" || { echo "profiling patch failed"; exit 1; }; \
echo "profiling patch applied to $entry"; \
fi
RUN --mount=type=secret,id=TURBO_API,env=TURBO_API \
--mount=type=secret,id=TURBO_TEAM,env=TURBO_TEAM \
+28
View File
@@ -0,0 +1,28 @@
# Performance findings
This file grew past 1,200 lines covering server memory, client rendering, methodology and a long
list of rejected ideas at once. It has been split by subject under [`docs/perf/`](./docs/perf/):
| document | subject |
| --- | --- |
| [docs/perf/README.md](./docs/perf/README.md) | Headline results and the change-by-change ratio table |
| [docs/perf/SERVER-MEMORY.md](./docs/perf/SERVER-MEMORY.md) | What the container's RAM is made of, layer by layer, and what reduced it |
| [docs/perf/CLIENT-PERFORMANCE.md](./docs/perf/CLIENT-PERFORMANCE.md) | Dashboard cost in Chrome: DOM, fibers, images, paint |
| [docs/perf/REJECTED.md](./docs/perf/REJECTED.md) | Ideas implemented, measured, and abandoned — with their numbers |
| [docs/perf/METHODOLOGY.md](./docs/perf/METHODOLOGY.md) | The tools, and the measurement traps that produced wrong answers |
## Result in one table
Server RAM, merge-base `06724cfeb` vs this branch, two alternated runs per side, cgroup **anonymous**
memory — the figure that counts against a container limit:
| stage | before | after | change |
| --- | --- | --- | --- |
| **peak** | 376.6 MiB | **279.3 MiB** | **−26%** |
| boot idle | 130.5 MiB | **62.4 MiB** | **−52%** |
| after sign-in | 247.4 MiB | **142.9 MiB** | **−42%** |
| dashboard loaded | 293.4 MiB | **170.4 MiB** | **−42%** |
| 4 tabs, 2 boards | 366.5 MiB | **258.2 MiB** | **−30%** |
| after 2 min idle soak | 323.7 MiB | **200.8 MiB** | **−38%** |
No before/after range overlaps at any stage. Cost: **+18% CPU**, solid and non-overlapping.
+1 -2
View File
@@ -13,7 +13,6 @@
# next.js
/.next/
/out/
/.source/
# production
/build
@@ -56,4 +55,4 @@ static/data/configs
# CI Stuff
screenshots
test-results
test-results
+9 -2
View File
@@ -1,3 +1,10 @@
{
"cSpell.words": ["Homarr", "Lidarr", "Nextcloud", "Radarr", "Readarr", "Sonarr"]
}
"cSpell.words": [
"Homarr",
"Lidarr",
"Nextcloud",
"Radarr",
"Readarr",
"Sonarr"
]
}
-18
View File
@@ -1,18 +0,0 @@
<!-- BEGIN:nextjs-agent-rules -->
# This is NOT the Next.js you know
This version has breaking changes — APIs, conventions, and file structure may all differ from your training data. Read the relevant guide in `node_modules/next/dist/docs/` (resolved from this file's directory; in monorepos the `next` package may not be visible from the repo root) before writing any code. Heed deprecation notices.
This block is written and re-added by `next dev` — verify at `node_modules/next/dist/server/lib/generate-agent-files.js`. Removing it from a diff only re-creates the uncommitted change; committing it with your work keeps the tree clean.
<!-- END:nextjs-agent-rules -->
## Fumadocs workflow
- Start with `README.md`, `src/lib/source.ts`, `source.config.ts`, and the matching route under `src/app/`.
- Read current [Fumadocs documentation](https://fumadocs.dev/docs) with Context7 before changing framework APIs. Check the installed version and types; latest examples may use a different search adapter.
- Preserve static export: search uses `staticGET` with the matching static client. Do not introduce a request-time search server into this app.
- Keep integration/widget metadata in the existing typed definitions. Check both rendered HTML and processed Markdown when changing MDX components; JSX alone does not guarantee readable search or AI output.
- Run the README validation commands after content or navigation changes. Search results and table-of-contents links must resolve to rendered anchors.
- Kapa is an independent hosted assistant. Keep its keyboard shortcut disabled so it does not override docs search; follow the README for its build configuration and crawler setup.
-1
View File
@@ -1 +0,0 @@
@AGENTS.md
-37
View File
@@ -1,37 +0,0 @@
# Homarr Documentation
## Register
product
## Users
People who run Homarr, administrators configuring self-hosted infrastructure, and contributors extending Homarr. They arrive with a concrete task such as installing Homarr, connecting a service, configuring a widget, managing users, or building a custom widget, and need an answer they can scan and apply quickly.
## Product Purpose
The documentation is the complete, authoritative operating guide for Homarr. It keeps setup steps, configuration contracts, screenshots, generated integration and widget metadata, API references, and Workshop tooling next to the application code. Success means a reader can find the right page quickly, understand the next action, copy reliable instructions, and verify the result without searching elsewhere.
## Brand Personality
Focused, dependable, and approachable. The writing is direct and practical. The interface should feel calm and familiar during long technical sessions while retaining Homarr's red identity.
## Anti-references
- Generic marketing layouts that replace Homarr's product identity with unrelated slogans or decorative conversion copy.
- Decorative dashboard aesthetics that compete with reading, including glass effects, large radii, card grids, and gratuitous motion.
- Generic generated-docs styling that obscures Homarr's identity or makes navigation feel detached from the main product.
- Dense reference portals that expose structure without giving readers a clear path through a task.
## Design Principles
1. Put the reader's task first: search, navigation, page outline, and the answer must remain easy to reach.
2. Preserve one source of truth: typed Homarr metadata and existing custom components stay connected to the monorepo.
3. Prefer familiar documentation patterns: the interface should disappear while a reader scans, copies, and follows instructions.
4. Make complete content feel manageable through strong hierarchy, progressive disclosure, and precise cross-links.
5. Treat machine readers as users through canonical Markdown, structured metadata, and standard LLM discovery endpoints.
6. Keep the root landing page recognizably Homarr: preserve its self-hosted dashboard promise and real product evidence while documentation routes remain task-first.
## Accessibility & Inclusion
Meet WCAG 2.2 AA for color contrast, keyboard operation, focus visibility, landmarks, headings, and interactive controls. Respect reduced-motion and system color-scheme preferences. Diagrams, screenshots, icons, and emoji must have meaningful text alternatives when they carry information.
+75 -54
View File
@@ -1,84 +1,105 @@
# Homarr documentation (`@homarr/docs`)
# Homarr Documentation (`@homarr/docs`)
The source for [homarr.dev](https://homarr.dev), built with Next.js and Fumadocs UI inside the Homarr monorepo.
The official [homarr.dev](https://homarr.dev) documentation site, built with [Docusaurus 3](https://docusaurus.io/) and living inside the Homarr monorepo.
## Develop
## Why it's in the monorepo
From the repository root:
Having documentation alongside the application code enables:
- **Two-way linking** — the docs app imports from `@homarr/definitions` for type-safe integration/widget metadata, and the main app references doc paths via generated sitemap types
- **Atomic PRs** — code changes and their documentation updates ship in the same pull request
- **Shared tooling** — same Node version, pnpm catalog, oxlint/oxfmt, CI pipeline
- **Build-time validation** — Docusaurus strict mode catches broken links and references at build time
## Development
From the **monorepo root**:
```bash
# Start the docs dev server (port 3003)
pnpm dev:docs
# Or from within this directory
pnpm dev
```
The site runs on `http://127.0.0.1:3003`. To use a local Workshop API:
## Build
```bash
WORKSHOP_API_URL=http://127.0.0.1:8090 pnpm dev:docs
# Build docs only
pnpm turbo build --filter=@homarr/docs
# Build everything (docs + main app + tasks + websocket)
pnpm build
```
## Validate
## Lint & Format
Uses oxlint and oxfmt (same as the rest of the monorepo):
```bash
pnpm --filter @homarr/docs typecheck
pnpm --filter @homarr/docs build
pnpm --filter @homarr/docs validate:links
pnpm --filter @homarr/docs verify:search
pnpm lint # oxlint
pnpm format # oxfmt (check)
pnpm format -- --write # oxfmt (fix)
```
The static export is written to `out/`. The build validates that every registered integration and widget has a docs
page. The link check validates internal Markdown and MDX links and rendered anchors against that export.
The search check exercises the exported index with title, heading, and body queries and verifies result destinations.
## Typecheck
## Write content
```bash
pnpm typecheck
```
Documentation lives in `docs/`; project posts live in `blog/`. Navigation is controlled by the nearest `meta.json`.
Integration and widget folders pair typed `index.ts` metadata with `index.mdx` content.
## Content Structure
Use standard Markdown and MDX. Shared components live in `src/components/`. Mermaid fences, tabs, callouts, zoomable
images, and syntax-highlighted code are supported by the Fumadocs pipeline.
```
docs/
├── getting-started/ # Installation guides, glossary, prerequisites
├── management/ # Boards, apps, users, integrations, settings, tasks, API
├── integrations/ # 56 integration guides (one folder each)
├── widgets/ # 43 widget guides (one folder each)
├── advanced/ # SSO, env vars, proxy, styling, CLI, development
└── community/ # FAQ, donate, license, translations, get-in-touch
```
When a user-facing change affects setup or behavior, update the matching page in the same pull request. See the
[documentation-sync skill](../../.agents/skills/documentation-sync/SKILL.md) for the code-to-content map.
Each integration/widget doc follows a consistent pattern:
## Published formats
- `index.ts` — typed metadata (`IntegrationDefinition` / `WidgetDefinition`)
- `index.mdx` — content using shared React components (`IntegrationHeader`, `WidgetHeader`, etc.)
- `/docs` — searchable documentation with typed navigation
- `/api-reference` — interactive reference generated from the OpenAPI schema
- `/llms.txt` — compact documentation and blog index for agents
- `/llms-full.txt` — complete processed Markdown corpus
- `/llms.mdx/docs/.../content.md` — raw Markdown for each docs page
- `/llms.mdx/blog/.../content.md` — raw Markdown for each blog post
- `/blog/rss.xml` — RSS feed for project posts
## Importing from homarr packages
Every docs page and blog post includes copy-Markdown and view-source actions. Search is generated at build time and
runs locally in the browser; it does not depend on an external crawler. Markdown exports and search resolve
integration credentials and widget defaults from the same typed metadata used by the pages.
The getting-started overview and complete Custom JSX component catalog are also included, with property types,
binding contracts, and blocked capabilities taken from the published catalog JSON.
The docs app has `@homarr/definitions` as a workspace dependency. You can import integration kinds, widget kinds, and other definitions directly:
Published Docusaurus `/docs/category/...` addresses remain available as static compatibility pages. They forward to
the corresponding overview or developer setup guide, preserving query strings and anchors, and provide a direct
link when JavaScript is disabled. Their canonical metadata points to the destination; compatibility pages are
excluded from search and the sitemap.
```typescript
import { IntegrationKind } from '@homarr/definitions';
```
## Kapa AI
This enables type-safe references to homarr domain concepts within documentation code.
The **Ask AI** launcher uses Homarr's existing public Kapa Website ID. The root layout loads the widget once across
client-side navigation. **Search** and `Ctrl+K` / `Cmd+K` always use Fumadocs search, including when Kapa is unavailable.
## Search (Algolia DocSearch)
Set `KAPA_WEBSITE_ID` at build time to use a different Kapa Website Widget integration. An explicitly empty value
disables the widget, for example `KAPA_WEBSITE_ID= pnpm --filter @homarr/docs build`. This is a public integration ID
included in exported pages, not a Kapa API key. Rebuild the static export after changing it.
Search uses the `Docusaurus` Algolia index (crawler `cd77a285-2756-4557-bf21-ee703748df15`). A separate `markdown` index is built for LLM markdown indexing.
In the Kapa dashboard, confirm the integration is live and enable the production and preview domains. Restoring the
script does not verify that the historical integration is still active. Follow the
[Kapa widget setup](https://docs.kapa.ai/integrations/website-widget/quickstart) for domain restrictions and CSP settings.
Crawler reference config: [`docsearch.config.js`](docsearch.config.js).
After changing the documentation structure, review the Website Crawl source in Kapa: use the deployed `/docs` URL as
the start URL and preview `main` as the content selector. Replace any old Docusaurus-specific selector and check that
titles, headings, code samples, and integration details appear in the extracted Markdown. Docs are exported as HTML;
the site also publishes `/sitemap.xml` and permits crawling in `/robots.txt`. The `/llms.txt` exports do not configure
Kapa ingestion automatically. See [Kapa source setup](https://docs.kapa.ai/getting-started/index-your-first-source).
**Verify search health:**
Before publishing, check that Ask AI opens after navigating between pages, ask a documentation question, and verify
its citations point to the current docs. In a separate check, block `widget.kapa.ai` and confirm search still works.
```bash
pnpm verify:search
```
## Tech Stack
| Layer | Technology |
|---|---|
| Site generator | Docusaurus 3 (`@docusaurus/preset-classic`) |
| Styling | Tailwind CSS 3 + Docusaurus Infima |
| Icons | `@tabler/icons-react` |
| Diagrams | Mermaid (`@docusaurus/theme-mermaid`) |
| Search | Algolia DocSearch |
| Analytics | PostHog |
| Charts | @nivo/line |
## Contributing
When making changes to the homarr codebase that affect user-facing behavior, **always update the corresponding documentation in this app**. See `.cursor/rules/documentation-sync.mdc` for the complete mapping of code changes to doc locations.
@@ -1,7 +1,6 @@
---
slug: documentation-migration
title: Migration of Documentation
date: 2022-06-22
authors:
- ajnart
- manuel-rw
@@ -13,7 +12,7 @@ The documentation has been re-written partly and includes now many animations, g
We are still figuring things out and highly depend on your feedback.
{/* Historical excerpt boundary. */}
<!-- truncate -->
## Why we migrated
@@ -33,4 +32,4 @@ After we've reviewed your changes, we'll merge your PR and this documentation wi
## TL;DR
The documentation on GitHub Wiki will soon disappear. It will be replaced with this Docusaurus-powered website. You'll find much more information here with higher detail and searchability.
The documentation on GitHub Wiki will soon disappear. It will be replaced with this Docusaurus-powered website. You'll find much more information here with higher detail and searchability.
@@ -1,7 +1,6 @@
---
slug: translations
title: Help us make Homarr multilingual
date: 2022-08-28
authors:
- ajnart
- manuel-rw
@@ -14,8 +13,8 @@ We've been working hard to offer Homarr in many languages. The next update of Ho
![](img/homarr-we-need-you.webp) <br/>
Yes, we need **YOU**! We want to make Homarr accessible to as many people as possible. Thus, we need to translate all content into many different languages. We need your help for that! [You can help us now to translate Homarr into as many languages as possible](/docs/community/translations).
{/* Historical excerpt boundary. */}
<!-- truncate -->
### About global translations / translations per user
We are aware that many people would like to configure translations globally or configure language per user. Currently, translations are stored in the Cookies, which will expire after some time (currently, it's set to 30 days). We will implement more granular language settings in future versions of Homarr, since it requires major changes to Homarr. Stay tuned! 😎
We are aware that many people would like to configure translations globally or configure language per user. Currently, translations are stored in the Cookies, which will expire after some time (currently, it's set to 30 days). We will implement more granular language settings in future versions of Homarr, since it requires major changes to Homarr. Stay tuned! 😎
@@ -1,6 +1,5 @@
---
title: "We present: Homarr 0.11 - the most complex Homarr version yet!"
date: 2023-01-11
authors:
- ajnart
- manuel-rw
@@ -21,7 +20,7 @@ It contains many breaking changes and huge improvements to your favourite dashbo
- Huge design and usability improvements to older components to Homarr
- Better handeling of exceptions and unexpected issues
{/* Historical excerpt boundary. */}
<!-- truncate -->
## Complete implementation of highly customizable drag and drop library "Gridstack"
@@ -1,6 +1,5 @@
---
title: "We present: Homarr 0.12 - The fastest homarr version ever!"
date: 2023-04-16
authors:
- ajnart
- manuel-rw
@@ -16,12 +15,14 @@ This applies to all widgets, but in the case of the IFrame widget, it enables yo
- Integrate monitoring tools like Grafana, Splunk or Elasticsearch
- Integrate your services (for example Sonarr, Radarr and Overseerr) directly into Homarr
![](https://media.discordapp.net/attachments/988335098493542410/1094665745226539168/update-teaser.png?width=1046&height=666)
## Edit mode password
We've added a setting to disable editing completely. This will allow you to restrict edit access without password-protection.
use the `EDIT_MODE_PASSWORD` ENV variable to set a password for edit mode.
turn on `DISABLE_EDIT_MODE=TRUE` to disable edit mode on startup.
{/* Historical excerpt boundary. */}
<!-- truncate -->
## 😎 Blazingly fast
We've implemented a new caching system, that will cache all data in your browser, meaning that Homarr will load within milliseconds, as if it was *powered by rust and blazingly fast* !!!
@@ -71,4 +72,4 @@ Now, you can use the following environment variables to configure this:
- ``DOCKER_HOST``
- ``DOCKER_PORT``
**Full Changelog**: https://github.com/ajnart/homarr/compare/v0.11.5...v0.12.0
**Full Changelog**: https://github.com/ajnart/homarr/compare/v0.11.5...v0.12.0
@@ -1,6 +1,5 @@
---
title: "Authentication in Homarr - 0.14 is here!"
date: 2023-11-10
authors:
- ajnart
- manuel-rw
@@ -24,7 +23,7 @@ provide you a viable solution today.
- 📝 Overhauled notepad widget with a complex graphical editor that enables easy editing right from your browser.
- :globe_with_meridians: Live Translation on Crowdin that enable you to directly translate using images to see what exactly you're translating
{/* Historical excerpt boundary. */}
<!-- truncate -->
## Breaking changes
@@ -84,4 +83,4 @@ We are aware that this update has many breaking changes and may confuse some use
For that reason, we updated our documentation and will enhance it further within the next few weeks.
We've also added this small button at the bottom of the page were you can ask **your questions** to an AI assistant:
![image](https://github.com/ajnart/homarr/assets/30572287/a56f721a-d022-42ad-813a-d85c89462a8c)
![image](https://github.com/ajnart/homarr/assets/30572287/a56f721a-d022-42ad-813a-d85c89462a8c)
@@ -1,6 +1,5 @@
---
title: "Updated documentation with better organization & design"
date: 2023-12-22
authors:
- manuel-rw
tags: [homarr, documentation, design, docs]
@@ -9,7 +8,7 @@ tags: [homarr, documentation, design, docs]
We'd like to let you know that we've updated our documentation and reorganized the pages.
This means, that some links may no longer be valid - we apologise if this causes any inconvenience for you.
{/* Historical excerpt boundary. */}
<!-- truncate -->
A few highlights include:
@@ -19,4 +18,4 @@ A few highlights include:
- Bug fixes for image zoom
- Added documentation for management pages and [new CLI tool](/docs/advanced/command-line/password-recovery/)
Please [let us know](/docs/community/get-in-touch) is something is not working as expected.
Please [let us know](/docs/community/get-in-touch) is something is not working as expected.
@@ -1,10 +1,9 @@
---
title: Homarr 1.0 - complete rewrite of all systems
date: 2024-09-23
authors:
- manuel-rw
---
# Homarr 1.0 - complete rewrite of all systems
We proudly present to you - Homarr ``1.0``.
We have started work on this new version back in 2023 and have worked tirelessly for months on this brand-new dashboard.
@@ -14,6 +13,8 @@ This will be a one time occurrence and version > ``1.0`` will have less breaking
This time more important than every; **read the breaking changes** list and **create a backup** before updating.
Some third party installation methods may take multiple days or weeks to update to the new version.
See [this guide on how to import your previous Homarr dashboard](/docs/getting-started/after-the-installation#importing-a-zip-from-version-before-100)
## Changes at a glance
- **Asynchronous integrations system**: In versions prior to ``1.0``, we fetched data upon each request synchronously.
This means that the user had to wait for the requests to finish so any data would appear.
@@ -29,7 +30,7 @@ reliability and scalability compared to < ``1.0``.
- **Improved performance**: We have upgraded the underlying framework of Homarr and take advantage
of an in-memory message bus and websockets to deliver data near realtime to you.
{/* Historical excerpt boundary. */}
<!-- truncate -->
## Breaking changes
:::warning
@@ -1,16 +1,16 @@
---
title: Open Beta 1.0
date: 2024-12-17
authors:
- meierschlumpf
---
# Open Beta 1.0
We are proud to announce the open beta of Homarr 1.0. This is a major milestone for us and we are excited to share it with you.
The open beta allows us to polish the last rough edges and gather feedback from you.
We are looking forward to your feedback and hope you enjoy the new features.
Let us know throgh discord or a [github issues](https://github.com/homarr-labs/homarr/issues/new/choose) if you encounter any issues.
{/* Historical excerpt boundary. */}
<!-- truncate -->
## How to use the beta with docker
@@ -36,4 +36,4 @@ services:
Then, run `docker compose up -d` in the same directory. This will start the Homarr container in the background.
Please note that the mounts have changed and it's not backwards compatible with versions before 1.0.0-beta.1.
You can find a list of all breaking changes [here](/blog/2024/09/23/version-1.0)
You can find a list of all breaking changes [here](/blog/2024/09/23/version-1.0)
@@ -1,20 +1,21 @@
---
title: Migrate from 0.15.10+ to 1.0.0
date: 2025-01-19
authors:
- meierschlumpf
---
# Migrate from 0.15.10+ to 1.0.0
In this guide, we will show you how to migrate from Homarr 0.15.10+ to 1.0.0.
If you are still on a version before 0.15.10, please update to 0.15.10 where we introduced the migration export page.
{/* Historical excerpt boundary. */}
<!-- truncate -->
## Preserve your old installation
## Export your data
Homarr 1.0 is a complete rewrite and cannot reuse the old Docker mounts. Keep the 0.15 installation and its data intact while you build the replacement. The historical `Management` > `Tools` > `Migrate to 1.0` export can still be kept as a record, but current Homarr releases cannot import that ZIP or its secret key.
Plan to recreate your apps, integrations, and boards manually in the new installation. A SQLite database backup can be restored from the onboarding welcome screen only when it was created by a compatible current Homarr release; it is not a conversion path for a 0.15 database.
As 1.0 is a complete rewrite of Homarr and therefor not compatible with the old docker mounts, we need to export your data first.
For this you can navigate to the `Migrate to 1.0` page which you can find under `Management` > `Tools` > `Migrate to 1.0`.
Here you can select which things you want to export and download the zip file.
It will also show you a secret key which you will need to import the secret data like user password hashes, salts and integration secrets.
![Export page](./img/old-homarr-export-data.png)
@@ -22,20 +23,33 @@ Plan to recreate your apps, integrations, and boards manually in the new install
We highly recommend to create a parallel installation of Homarr 1.0.0.
So if something with for example the import does not quite work, the data in the old installation is still available.
To do this you can follow the appropriate [installation guide](/docs/getting-started) for your system.
To do this you can follow the appropriate [installation guide](https://homarr.dev/docs/category/installation-1) for your system.
Please note that if you are using docker you need to specify a diffrent external port for the new installation as the old one had the same port.
So in your compose file the ports would be `7576:7575` instead of `7575:7575`.
## Import your data
Once the installation is completed you can access the new installation. During the onboarding you'll need to select `Import from Homarr before 1.0`.
:::warning Legacy guide
![Onboarding initial step](./img/onboarding-step-1.png)
The pre-1.0 ZIP importer is no longer included in current Homarr releases. Keep the old installation available until you have manually verified the replacement. Compatible SQLite backups from current releases can still be restored from the onboarding welcome screen.
After that you can upload the zip file you downloaded in the first step.
:::
![Onboarding import dropzone](./img/old-homarr-import-dropzone.png)
After that you'll get a list with all your boards that can be imported. You can configure how it should be imported and can select which screensizes it should import.
![Onboarding import list](./img/old-homarr-import-choose-imported-data.png)
Once you have selected everything you can click on `Confirm import and continue` which will open the encryption key input dialog where you can input the secret key you got during the export.
![Onboarding import key input](./img/old-homarr-import-decryption-token.png)
After that you can continue the onboarding and you should see all your boards and data in the new installation.
See [After the installation](https://homarr.dev/docs/getting-started/after-the-installation) for more information.
## Remove the old installation
After you are happy with the new installation and everything is working as expected you can remove the old installation.
Please make sure to backup the data before you remove the old installation.
Also once the port 7575 is free again you can change the port mapping of the new installation to `7575:7575` if you want to.
Also once the port 7575 is free again you can change the port mapping of the new installation to `7575:7575` if you want to.
@@ -1,14 +1,14 @@
---
title: How to use Argus with Homarr
date: 2025-08-02
authors:
- meierschlumpf
---
# How to use Argus with Homarr
In `v1.32.0` we added the endpoint `/api/info` which returns the current version of Homarr and can therefore be used in Argus to watch for updates.
This guide will show you how to set up a user with apiKey to access the endpoint without providing admin access through it.
{/* Historical excerpt boundary. */}
<!-- truncate -->
![Argus](./img/argus.png)
@@ -64,4 +64,4 @@ Now you can configure Argus to use the user and apiKey we created.
![Argus configuration](./img/argus-configuration.png)
Replace the `<secret>` with the apiKey you created and the deployed version url with the url of your Homarr instance.
Replace the `<secret>` with the apiKey you created and the deployed version url with the url of your Homarr instance.
Binary file not shown.

Before

Width:  |  Height:  |  Size: 79 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 33 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 58 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 95 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 11 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 418 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 26 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 25 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 26 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 618 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 28 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 23 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 24 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 40 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 28 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 36 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 32 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 25 KiB

@@ -1,169 +0,0 @@
---
title: Homarr 2.0.0 - Workshop, Custom Widgets and a rebuilt board
date: 2026-09-03
authors:
- ajnart
tags: [homarr, update, version, assistant, custom widgets, dashboards, docker]
description: Homarr 2.0 introduces Custom Widgets and Workshop, rebuilt board editing, Assistant tooling, responsive layouts, onboarding, and more.
image: ./img/homarr-v2-dashboard.webp
draft: false
---
# Homarr 2.0.0 - Workshop, Custom Widgets and a rebuilt board
First, thank you.
Many of you used the v2 demo when it was still rough, found bugs and came back to test the next build. Others translated Homarr, answered questions, contributed code or simply kept using it. We would not be shipping this without you.
We are also planning a small batch of **free, limited-edition Homarr swag** for Workshop contributors and a few randomly picked community members. It is not for sale and it is not for profit. No Homarr merch empire, sorry.
Homarr 2.0 is an upgrade, not a brand-new app wearing the same logo. Here is what actually changed.
![A populated Homarr 2.0 dashboard with compact widgets, an Assistant, and an application rail](./img/homarr-v2-dashboard.webp)
{/* Historical excerpt boundary. */}
:::warning Back up before upgrading
Database migrations run automatically when Homarr starts. Create a
[Homarr backup](/docs/management/backup) first. Do not test a prerelease image on your only production volume.
Backups are boring. Restores are worse.
:::
## Custom Widgets v2 and Workshop
### Build almost any widget
**This is the big one.** We think it is the most important Homarr update we have made.
Custom Widgets v2 lets you build widgets from JSX, API requests, actions and typed settings. Write one yourself or describe it to Assistant, then test it in the workbench. We think this is enough to make almost any dashboard widget—even that oddly specific component only three people need. Those three people no longer have to wait for us.
This was partly inspired by [Google's upcoming Create My Widget feature](https://blog.google/products-and-platforms/platforms/android/gemini-intelligence/): describe a widget in normal language and let AI make it. Coding knowledge is optional. Knowing what the API does still helps; AI has not fixed bad documentation. Yet.
![The Custom Widget v2 workbench after a live exchange-rate request successfully renders its preview](./img/custom-widget-workbench.webp)
A `homarr-custom-widget-v2` definition can use several fixed-origin API sources. Credentials are encrypted separately and stay out of exports and Workshop submissions. Actions support `GET`, `POST`, `PUT`, `PATCH` and `DELETE`; requests have host checks, timeouts, size limits, permissions and rate limits. Custom Widgets remain **Beta** in 2.0.
### Share widgets and Custom CSS
**Workshop is the other half.** Publish and install Custom Widgets, then update, vote on, comment on and report submissions. Workshop is not only for widgets: you can share your **Custom CSS** there too. Deployment URLs and credentials stay on your Homarr instance.
![The Community Workshop catalog inside Custom Widget management](./img/community-workshop.webp)
:::warning Workshop is moderated
AI makes widgets easy to produce. It does not make the first thing it coughs up worth publishing. **Do not dump untested slop into Workshop.** Test it, add a useful screenshot and explain its setup. Workshop is moderated.
:::
### The ultimate Homarr update
We have called this the **"ultimate Homarr update"**. Not because work stops, but because every new service or widget idea no longer needs a full Homarr release; the community can build and share it instead.
## Board editing, rebuilt
### Drag and drop
We threw out the old drag-and-drop system and rebuilt it on [`dnd-kit`](https://dndkit.com/). Moves and eight-direction resizes are transactional: Homarr previews the result, checks collisions and only saves a valid layout. Bad moves roll back instead of wrecking the board. In short: less fighting with your own dashboard.
![A Calendar tile moving across the rebuilt Homarr 2.0 board grid](./img/drag-and-drop.gif)
### Goodbye Groups and sections, bonjour Containers
Groups and dynamic sections did nearly the same job, so **Containers replace both**. They hold apps, widgets and other Containers. Hold <kbd>Ctrl</kbd> or <kbd>Command</kbd> to move several items at once.
### One board, several layouts
Boards also have Base and Mobile layouts, plus optional breakpoints. **Reset from Base** creates a Mobile starting point without duplicating the board.
## Homarr Assistant
### Work with live Homarr data
Assistant reads live Homarr data and has tools for boards, apps, integrations, widgets, Docker and supported services. It can answer questions, make approved changes and build Custom Widgets. It follows your permissions: reads can run automatically, while changes require approval by default.
![The Homarr Assistant expanded over a live dashboard with health, exploration, media, and board actions](./img/assistant-control-surface.webp)
Open it with <kbd>Shift</kbd> + <kbd>A</kbd>. <kbd>Cmd</kbd>/<kbd>Ctrl</kbd> + <kbd>/</kbd> searches it.
### The Homarr provider
:::note
Assistant ships with a **"Homarr provider"**. It is free, will never offer paid plans, and is paid for by [Ajnart](https://github.com/ajnart) using the ad revenue from this website.
It only uses zero-data-retention (ZDR) providers. [The provider is open source](https://github.com/homarr-labs/homarr/blob/release/v2/apps/workshop/homarr_provider.go), and neither prompts nor responses are logged. We are not selling your data. Pinky promise.
:::
### External assistants through MCP
External assistants can use the same permission-checked tools through `/api/mcp`, with an API key or OAuth 2.1 with PKCE.
## The rest (there is a lot)
### Brand-new onboarding
A six-stage setup studio handles the first admin, core settings, Docker discovery, apps, integrations and the first board.
![The Homarr 2.0 setup studio on its Essentials step, with all six setup stages shown](./img/onboarding-studio.webp)
### Board switcher
Boards now live in a searchable gallery with keyboard navigation. Open it from the header or press <kbd>Shift</kbd> + <kbd>C</kbd>.
![A visual board card inside the searchable Homarr board switcher](./img/board-switcher.webp)
### Command menu and keybinds
<kbd>Cmd</kbd>/<kbd>Ctrl</kbd> + <kbd>K</kbd> finds resources, pages, settings, users, groups and commands. More
keybinds cover Assistant, editing, selection and advanced widget views.
### Customizable header
Pick which buttons appear, put them on the left, center or right, and change their order.
![The Header studio showing left, center, and right zones plus available controls](./img/header-studio.webp)
### Instance branding
Change the name, logo, favicon, colors, corner styles and login page. Global CSS applies instance-wide; board CSS can override it. Only give that power to admins you trust. CSS remains CSS.
![Instance branding controls for application identity, palette, corner style, and global CSS](./img/instance-branding.webp)
### Widgets
Add content without leaving the board. Settings have live previews, and supported widgets have advanced views. Weather and Clock were redesigned; Air Quality, Countdown and Timer are new.
### Docker and Podman
`DOCKER_ENDPOINTS` supports several hosts. Assisted setup and container labels can create apps, integrations, Containers, widgets and health checks.
![Assisted Docker setup listing discovered services that can become apps or integrations](./img/docker-assisted-setup.webp)
Homepage-compatible labels work too.
### Permissions
Groups use a new Viewer, Editor and Admin matrix. The API, Assistant and MCP follow the same rules. LDAP can create users on first sign-in; OIDC can link verified accounts and synchronize groups.
### Performance
Board startup work runs in parallel, heavy screens load when opened, and identical integration requests share work instead of doing it twice.
## Before upgrading
- Keep a backup and use the final release image, not a prerelease tag.
- Legacy Custom JSX definitions enter a migration state. Homarr keeps encrypted credentials and an archived copy until the v2 definition validates.
- Mobile layouts place rail content in the main flow instead of displaying rails.
- Custom Widgets remain Beta.
- Assistant and MCP do not bypass permissions.
## Links
- Try the [Homarr 2.0 preview](https://app-v2.preview.homarr.dev).
- Follow the [v2 beta setup guide](/docs/getting-started/upgrade-v2).
- Follow the [v2 release pull request](https://github.com/homarr-labs/homarr/pull/6545) and [public beta thread](https://github.com/homarr-labs/homarr/issues/6600).
- Compare [v1.76.2 with the release/v2 branch](https://github.com/homarr-labs/homarr/compare/v1.76.2...release/v2).
- Read the guides for [Boards](/docs/management/boards), [Custom Widgets](/docs/management/custom-widgets), [Assistant](/docs/management/assistant), [MCP](/docs/management/mcp) and [Docker labels](/docs/integrations/docker-labels).
-24
View File
@@ -1,24 +0,0 @@
{
"$schema": "https://ui.shadcn.com/schema.json",
"style": "base-nova",
"rsc": false,
"tsx": true,
"tailwind": {
"config": "",
"css": "src/app/global.css",
"baseColor": "mist",
"cssVariables": true,
"prefix": ""
},
"aliases": {
"components": "@/components",
"utils": "@/lib/utils",
"ui": "@/components/ui",
"lib": "@/lib",
"hooks": "@/hooks"
},
"iconLibrary": "tabler",
"rtl": false,
"menuColor": "default",
"menuAccent": "subtle"
}
+9
View File
@@ -0,0 +1,9 @@
{
"label": "Advanced",
"collapsed": true,
"position": 4,
"link": {
"type": "generated-index",
"title": "Advanced Concepts or features"
}
}
@@ -0,0 +1,4 @@
{
"collapsed": true,
"label": "Command line interface"
}
@@ -1,180 +1,41 @@
# Developer Docker CLI
The host-side developer CLI builds and runs local Homarr images and pull-request previews with non-blocking background operations. Run it directly from a checkout with `pnpm dev:cli -- ...`; no shell-specific setup script is required. It is separate from the recovery CLI included in production images.
The host-side developer CLI builds and runs local Homarr images and pull-request previews. Run it directly from a checkout with `pnpm dev:cli -- ...`; no shell-specific setup script is required. It is separate from the recovery CLI included in production images.
To install an optional `homarr` binary, run `pnpm dev:cli:install` and make your Go bin directory available on `PATH`. GitHub CLI authentication (`gh auth login`) is only needed for pull-request features.
To install an optional `homarr` binary, run `pnpm dev:cli:install` and make your Go bin directory available on `PATH`. GitHub CLI authentication is only needed for pull-request features.
## Requirements
- Docker or OrbStack
- Go 1.25+ (when building the CLI from source)
- GitHub CLI authenticated with `gh auth login` (for pull-request and CI features)
- A [Nerd Font](https://www.nerdfonts.com/) enabled in your terminal (recommended for icons, status indicators, and TUI glyphs)
## Interactive Development Browser
## Browse development images
```bash
pnpm dev:cli --
# or with global binary:
homarr
```
The main browser combines local `homarr:*` images with open pull requests and their GHCR preview status. Select an image to start or stop it, use `R` to build a remote pull request locally in the background or rebuild a local image, or use `p` to force a remote preview pull. Press `d` (or `1`/`2`) to toggle between the image browser and the instance dashboard.
The main screen combines local `homarr:*` images with open pull requests and their GHCR preview status. Select an image to start or stop it, use `R` to build a remote pull request locally in the background or rebuild a local image, or use `p` to force a remote preview pull. Press `d` to switch to the instance dashboard and `d` again to return without losing the browser state. The explicit `dev` and `dash` subcommands start on either screen.
A locally built PR image is folded into its remote PR row only when its recorded revision matches the current GitHub head SHA (marked `◆ local`). When the matching GHCR image is also available, pressing `Enter` opens a local/remote selector with local preselected. Outdated local builds remain separate rows.
A locally built PR image is folded into its remote PR row only when its recorded revision exactly matches the current GitHub head SHA. The row is marked `◆ local`, and starting it prefers that local image without a registry pull. Outdated local builds remain separate rows. Navigation and dashboard switching remain available during background builds.
### Multi-Source Sidebar
Press `Tab` to cycle between sidebar tabs:
- **`[logs]`**: Live container output streams (press `f` to follow/pause, `L` to pin to the selected container, `Page Up`/`Page Down` to scroll).
- **`[build]`**: Real-time BuildKit step streaming for active local builds.
- **`[ci]`**: GitHub Actions CI check statuses, run duration, and workflow links for pull requests.
### Background Task Overlay
Pulls and builds run asynchronously in the background. Press `t` to open the tasks overlay, showing:
- Layer-by-layer download progress and transfer speeds (via Docker Engine API).
- Active build stages and logs.
- Task cancellation and status.
### Browser Shortcuts
| Key | Action |
| ---------------------- | --------------------------------------------------------------------------------------------- |
| `/` | Filter rows by PR number, title, author, branch, CI, image, state, or port |
| `↑` / `↓` or `j` / `k` | Navigate rows |
| `Enter` or `Space` | Start the available image; choose local or remote when both exist; stop if running |
| `R` | Build locally in background (or rebuild local image) |
| `p` | Pull remote PR image in background and start container |
| `m` | Manage mode (`b` rebuild, `i` delete image, `d` delete data, `c` remove container, `p` prune) |
| `t` | Toggle background tasks overlay |
| `Tab` | Cycle sidebar tabs (`[logs]`, `[build]`, `[ci]`) |
| `l` / `L` | Toggle sidebar visibility / Pin logs to selected instance |
| `f` | Follow or pause logs |
| `M` | Toggle demo mode for next launch |
| `b` | Toggle bot pull requests |
| `o` / `a` | Open PR on GitHub / Open running instance in browser |
| `c` | Copy instance URL to clipboard |
| `r` | Refresh data |
| `1` / `2` / `d` | Switch screen (development browser / instance dashboard) |
| `?` / `q` | Help overlay / Quit (cancels active background tasks) |
## Instance Dashboard
## Build images
```bash
homarr dash
# or
pnpm dev:cli -- dash
pnpm dev:cli -- build <name>
pnpm dev:cli -- build --pr <number> [name]
```
The instance dashboard displays all running and stopped Homarr containers, published ports, health states, and resource volumes.
The first command builds the current checkout as `homarr:<name>`. The pull-request form uses a depth-one clone in a temporary directory and defaults to `homarr:pr-<number>` when `name` is omitted.
| Key | Action |
| ------------------ | ---------------------------------------- |
| `Enter` or `Space` | Start or stop selected instance |
| `s` / `S` | Stop / Restart instance |
| `x` | Remove container (preserves data volume) |
| `m` | Manage mode for selected instance |
| `R` | Rebuild from recorded build provenance |
## CLI Commands
The commands below use the installed `homarr` binary. Without global installation, prefix each command with `pnpm dev:cli --` (for example, `pnpm dev:cli -- list`).
### Launch Instances
## Rebuild images
```bash
homarr run dev
homarr run --detach dev
homarr run --pr 6441
homarr run --pr 6441 --demo
homarr --pr 6441 --env WORKSHOP_WEB_URL=https://app-v2.preview.homarr.dev/
pnpm dev:cli -- rebuild <name>
```
- `--pr <number>`: Pull and run a PR image from GHCR (shorthand: `homarr --pr <number>`).
- `--demo` (`-m`): Start with mock integrations and sample data.
- `--detach`: Start container in background.
- `--env` (`-e`): Pass additional environment variables (repeatable: `--env KEY=VALUE`).
Images built by this CLI record their source checkout, revision, and optional pull-request number. Rebuild uses that provenance to select the original checkout or create a fresh temporary pull-request checkout.
### Build & Rebuild
## Run images
```bash
# Build current checkout as homarr:<name>
homarr build feature
# Build a pull request locally in a temporary checkout
homarr build --pr 6441
# Rebuild an existing image from recorded provenance
homarr rebuild feature
homarr run <name>
homarr run --pr <number>
```
Images built with the CLI record their source checkout, revision SHA, and optional PR number. `homarr rebuild` reads this provenance to reconstruct the build without manual configuration.
### GitHub CI Checks
```bash
# View CI checks for a specific PR
homarr ci 6441
# View CI checks for the current branch
homarr ci
# Watch CI checks until completion
homarr ci 6441 --watch
```
### Interactive Shell
```bash
# Open interactive shell in running container
homarr shell homarr-dev
# Run specific command in container
homarr shell homarr-dev pnpm db:migration:sqlite:run
```
### Image & Volume Management
```bash
# List local images with build provenance (source, revision, PR)
homarr images
# Delete local image
homarr images rm feature
# List data volumes with owner PR/tag and in-use status
homarr data
# Preview deleting a data volume and its owning container
homarr data rm homarr_feature_data
# Permanently delete the data after reviewing the preview
homarr data rm homarr_feature_data --force
# Remove stopped containers (prompts without --force)
homarr prune --force
```
Forced data deletion removes the derived owner container only when it mounts the exact selected volume. A same-named container that does not own that volume stops the operation.
### Instance Operations & Diagnostics
```bash
# List instances in a script-friendly table
homarr list
# Follow container logs
homarr logs homarr-dev
# Stop, restart, remove, or open container in browser
homarr stop homarr-dev
homarr restart homarr-dev
homarr remove homarr-dev
homarr open homarr-dev
# Verify Docker daemon, Engine API, GHCR access, and gh CLI authentication
homarr doctor
```
Add `--detach` to run in the background, `--demo` to enable demo data, or repeat `--env KEY=VALUE` to provide additional environment variables.
@@ -14,4 +14,4 @@ homarr fix-usernames
This will convert all credential usernames to lowercase and update the database accordingly.
![Terminal output confirming credential usernames were converted to lowercase](./img/screenshot-cli.png)
![](./img/screenshot-cli.png)
Binary file not shown.

Before

Width:  |  Height:  |  Size: 768 KiB

+18 -62
View File
@@ -1,71 +1,27 @@
# Command line interface
The Homarr image includes a recovery CLI for operations that cannot be completed in the web interface. It is not an
application API; use the [API](/docs/management/api) or [MCP endpoint](/docs/management/mcp) for automation.
The Homarr image ships with a built-in recovery tool. This can be useful, if you do not have access to Homarr or if something is not working.
Please note that the CLI is an *emergency tool*, not an API tool.
Back up the database and stop Homarr before running a command that changes data.
Failed commands print an error and return a non-zero exit code, so shell scripts can stop safely.
:::caution
## Run the recovery CLI
Using the recovery CLi tool can cause problems when requests are being sent to Homarr while the tool is running. \
You should also create a backup of the database, in case you break something.
Open a shell in the container, then pass a command to `homarr`:
:::
```sh
docker exec -it <container-id> /bin/bash
homarr <command>
```
## Execute CLI with Docker
### QNAP Container Station
1. Execute ``docker ps | grep homarr`` on the root system with Docker to find the container ID of your Homarr container.
2. Run ``docker exec -it <container-id> homarr <your-command-goes-here>``. \
Replace ``<container-id>`` with your container ID from step 1. \
Replace ``<your-command-goes-here>`` with the command you want to execute (see sub-pages).
1. Open **Container Station**.
## Execute CLI with QNAP
![Container Station in the QNAP application menu](./img/334116069-e7ec4726-3df7-4dc6-a3e0-b745e7eef17f.png)
2. Select **Containers**.
![Containers menu in QNAP Container Station](./img/334116205-aa09ce40-0701-4f7d-b9f9-b431b9d207c5.png)
3. Select the Homarr container.
![Homarr container selected in QNAP Container Station](./img/334116303-714b5e07-b9de-4ec0-a2bc-def7ee38750f.png)
4. Select **Execute**.
![Execute action for the Homarr container in QNAP Container Station](./img/334116411-df1482f9-3aa9-4686-b77f-e0dca09e60b2.png)
5. Select `/bin/bash`, then select **Execute**.
![Bash selected in the QNAP execute console](./img/334116518-e1ddb515-99c7-4159-81bf-76e889d1e808.png)
6. Run `homarr <command>` in the console.
![Bash console open for Homarr in QNAP Container Station](./img/334116642-08c4f204-5cd0-4e4e-a785-0142e1b279ca.png)
See the pages in this section for available commands.
## Build a local branch harness
The repository's host-side CLI can build the current checkout and run it as a named Docker harness:
![Homarr developer CLI showing pull requests, local images, and PR CI status](./img/homarr-dev-cli.png)
```sh
pnpm cli harness setup feature-board
```
This creates `homarr:feature-board`, starts it with persistent demo data, and prints the assigned URL. Query the port
again with:
```sh
pnpm cli harness port feature-board
```
Pass environment overrides with repeated `--env` options:
```sh
pnpm cli harness setup feature-board \
--env WORKSHOP_WEB_URL=https://example.test \
--env FEATURE_FLAG=true
```
Related commands are `harness build <name>`, `harness run <name>`, and `harness stop <name>`.
1. Open Container Station in the QNAP Web Interface ![](img/334116069-e7ec4726-3df7-4dc6-a3e0-b745e7eef17f.png)
2. Select Containers in the menu bar on the left ![](img/334116205-aa09ce40-0701-4f7d-b9f9-b431b9d207c5.png)
3. Select your Homarr Container ![](img/334116303-714b5e07-b9de-4ec0-a2bc-def7ee38750f.png)
4. Click "Execute" in the top right of the Window. ![](img/334116411-df1482f9-3aa9-4686-b77f-e0dca09e60b2.png)
5. Select /bin/bash and click Execute ![](img/334116518-e1ddb515-99c7-4159-81bf-76e889d1e808.png)
6. Run the Homarr command and press enter (eg. homarr reset-owner-password) ![](img/334116642-08c4f204-5cd0-4e4e-a785-0142e1b279ca.png)
@@ -7,7 +7,6 @@ This guide requires that you understand [how to access the CLI](../).
:::
Using these commands, you can manage the integrations from the CLI:
```bash
Group of commands to manage integrations
@@ -44,7 +43,7 @@ Flags:
## Delete Integration
- Requires a database connection
- Requires one parameter: `id` or `name`
- Requires one parameter: ``id`` or ``name``
- Deletes an integration permanently from the database
```bash
@@ -58,4 +57,4 @@ Flags:
-n, --name string Name of the integration
-h, --help help for integrations delete
-v, --version version for homarr-cli
```
```
@@ -1,4 +0,0 @@
{
"title": "Command line interface",
"pages": ["index", "users", "integrations", "password-recovery", "fix-usernames", "development"]
}
Loaded 100 of 2753 files, more files were not shown because too many files have changed in this diff. Show more