feat(plugins): add llama.cpp server integration and widget (#6765)

apps/docs/docs/integrations/llama-cpp/index.mdx

@@ -0,0 +1,55 @@
+---
+title: "llama.cpp"
+description: "Connect Homarr to a llama.cpp llama-server to monitor your local LLM: health, loaded model, generation speed and request status."
+hide_title: true
+---
+
+import { IntegrationHeader } from "@site/src/components/integrations/header";
+import { IntegrationCapabilites } from "@site/src/components/integrations/widgets";
+import { AddingIntegration } from "@site/src/components/integrations/adding";
+import { IntegrationSecrets } from "@site/src/components/integrations/secrets";
+import { llamacppIntegration } from ".";
+import { llamacppWidget } from "@site/docs/widgets/llama-cpp";
+
+<IntegrationHeader integration={llamacppIntegration} categories={["System Monitoring"]} />
+
+The llama.cpp integration lets Homarr connect to a [llama-server](https://github.com/ggml-org/llama.cpp) instance so you can see the health of your local LLM, which model is loaded, and how fast it is generating — all from your dashboard.
+
+### Widgets & Capabilities
+
+<IntegrationCapabilites items={[{ widget: llamacppWidget }]} />
+
+### Adding the integration
+
+<AddingIntegration />
+
+:::info URL configuration
+Use the root URL of your `llama-server`, for example `http://192.168.1.50:8080`. Homarr calls the `/health`, `/v1/models`, `/metrics` and `/slots` endpoints under that URL automatically.
+:::
+
+:::tip Metrics
+Generation speed and token counters are read from the Prometheus endpoint at `/metrics`. Start `llama-server` with the `--metrics` flag so these values are available:
+
+```shell
+llama-server --model /path/to/model.gguf --metrics
+```
+
+Without `--metrics` the widget still shows health, model information, and active-request speed from `/slots`; metrics-backed counters and aggregate speed stay empty.
+:::
+
+:::info Context and per-request speed
+Context (KV cache) usage and the in-flight request data are read from the `/slots` endpoint, which is available on every recent llama-server build and requires no extra flags. The per-request generation speed shown while a request is running is derived by tracking the request's decoded token count across polls, so it reflects the average speed of the request that is currently being generated.
+:::
+
+### Secrets
+
+llama-server does not require authentication by default. The integration only needs the URL of your server.
+
+<IntegrationSecrets
+  secrets={[
+    {
+      credentials: [],
+      steps: [],
+    },
+  ]}
+/>

apps/docs/docs/integrations/llama-cpp/index.ts

@@ -0,0 +1,9 @@
+import { IntegrationDefinition } from "@site/src/types";
+
+export const llamacppIntegration: IntegrationDefinition = {
+  name: "llama.cpp",
+  description:
+    "Connect to a local llama.cpp llama-server to show model, throughput and request status on your dashboard.",
+  iconUrl: "https://cdn.jsdelivr.net/gh/selfhst/icons/svg/llama-cpp.svg",
+  path: "../../integrations/llama-cpp",
+};

apps/docs/docs/widgets/llama-cpp/index.mdx

@@ -0,0 +1,43 @@
+---
+title: "llama.cpp"
+description: "Displays the health, loaded model and generation speed of a local llama.cpp llama-server."
+hide_title: true
+---
+
+import { WidgetHeader } from "@site/src/components/widgets/header";
+import { AddingWidget } from "@site/src/components/widgets/adding";
+import { WidgetConfig } from "@site/src/components/widgets/configuration";
+import { WidgetIntegrations } from "@site/src/components/widgets/integrations";
+import { llamacppWidget } from ".";
+import { llamacppIntegration } from "@site/docs/integrations/llama-cpp";
+
+<WidgetHeader widget={llamacppWidget} categories={["System Monitoring"]} />
+
+This widget shows the live status of your local [llama.cpp](https://github.com/ggml-org/llama.cpp) `llama-server`: whether it is online, which model is loaded (with quantization, context size, file size and parameter count), the generation speed in tokens per second, and how many requests are being processed. When the server is started with `--metrics`, it also reports total token counts and the prompt cache hit rate.
+
+### Generation speed
+
+The widget picks the most meaningful speed for the current state (hover the speed for an explanation of which one is shown):
+
+- **While a request is being generated** — the average speed of that request since it started, derived from the token count reported by `/slots`.
+- **While the server is busy but the per-request data is unavailable** — the live speed from the `/metrics` rate gauge.
+- **While the server is idle** — the average generation speed since the server started (cumulative tokens ÷ seconds), so the value stays meaningful instead of dropping to zero.
+
+### Supported Integrations
+
+<WidgetIntegrations
+  items={[
+    {
+      integration: llamacppIntegration,
+      note: "Displays health, model details, generation speed and request status",
+    },
+  ]}
+/>
+
+### Adding the widget
+
+<AddingWidget />
+
+### Configuration
+
+<WidgetConfig configuration={llamacppWidget.configuration} />

apps/docs/docs/widgets/llama-cpp/index.ts

@@ -0,0 +1,37 @@
+import { WidgetDefinition } from "@site/src/types";
+import { IconCpu } from "@tabler/icons-react";
+
+export const llamacppWidget: WidgetDefinition = {
+  icon: IconCpu,
+  name: "llama.cpp",
+  description: "Shows the health, loaded model and generation speed of a local llama.cpp llama-server.",
+  path: "../../widgets/llama-cpp",
+  configuration: {
+    items: [
+      {
+        name: "Show title",
+        description: "Displays the llama.cpp label and status badge at the top of the widget",
+        values: { type: "boolean" },
+        defaultValue: "yes",
+      },
+      {
+        name: "Show model info",
+        description: "Displays the loaded model name, quantization, context size and file size",
+        values: { type: "boolean" },
+        defaultValue: "yes",
+      },
+      {
+        name: "Show cache hit rate",
+        description: "Displays the prompt cache hit rate as a percentage bar",
+        values: { type: "boolean" },
+        defaultValue: "yes",
+      },
+      {
+        name: "Show context usage",
+        description: "Displays the context (KV cache) token usage as a progress bar",
+        values: { type: "boolean" },
+        defaultValue: "yes",
+      },
+    ],
+  },
+};