10000ge10000/luci-app-openclaw

luci‑app‑openclaw – OpenWrt UI for the OpenClaw AI gateway

What it is

• A LuCI (the web UI used by OpenWrt routers) plugin that lets you install, configure and run the OpenClaw AI gateway directly on a router.
• OpenClaw itself is a lightweight AI‑gateway that can host LLM models and expose them via chat channels (e.g., WeChat). This plugin provides the graphical front‑end for managing that service.

Why it matters

• Turns a home router into an on‑premise AI inference node without needing a separate server.
• Supports both x86_64 and ARM64 routers that use the musl C library, with automatic download and verification of a pre‑built Node.js 22 runtime.
• Includes built‑in support for the Tencent WeChat channel, making it easy to interact with the AI via a familiar messaging app.

Key features (as described in the README)

• One‑click environment setup – a “Install runtime” button in LuCI downloads and verifies Node.js and OpenClaw, then starts the service.
• Model management – a three‑tier model selector:
  1. Featured models (hard‑coded presets for each provider).
  2. Live list fetched from OpenClaw (openclaw models list).
  3. Manual entry for any model ID.
• WeChat channel wizard – installs the @tencent‑weixin/openclaw‑weixin plugin, creates the required system user, checks Python3, and guides the QR‑code login flow.
• Configurable install path – via the standard UCI variable openclaw.main.install_path (defaults to /mnt/data/openclaw).
• Robust CI – GitHub Actions build the Node.js musl binary, package the plugin as both .run self‑extracting installer and OpenWrt .ipk, and publish releases automatically.
• Extensive sanity checks – before installing it validates storage space, overlay writeability, TLS connectivity to WeChat endpoints, and permission of OpenClaw directories.

Installation options

1. Self‑extracting .run – download the latest release, run the script; no OpenWrt SDK needed.
2. OpenWrt package .ipk – download the .ipk and install with opkg.
3. Compile into a custom firmware – add the repo as a feed in feeds.conf.default, run scripts/feeds, select LuCI → Applications → luci‑app‑openclaw in make menuconfig, then compile.

First‑time use

1. Open LuCI → Services → OpenClaw → click Install runtime.
2. After the service starts, go to Web console to add an AI model and its API key.
3. Use the Configuration wizard to enable a messaging channel (e.g., WeChat).
4. Test by sending a message from the chosen channel; the gateway should reply using the selected LLM.

System requirements

Requirement	Detail
CPU	x86_64 or aarch64 (ARM64)
C library	musl (required for the pre‑built Node.js binary)
Dependencies	luci-compat, luci-base, curl, openssl-util, tar, script-utils
Storage	≥ 2 GB free
RAM	Recommended ≥ 1 GB

Development & contribution

• Source layout follows the typical OpenWrt package structure (Makefile, luasrc/, root/ etc.).
• Tests are provided under tests/ and run via sh tests/run_all.sh.
• Contributions are welcomed via Issues and Pull Requests.

License – GPL‑3.0.

---

All information above is taken directly from the repository’s README; no additional features have been inferred.