LightRAG Knowledge Base

Raycast extension for your self-hosted LightRAG server: query a graph-based RAG knowledge base, inspect documents and pipeline status, explore the knowledge graph, and use Raycast AI tools against your own data.

You run LightRAG on your machine or network; this extension only connects to your server URL and credentials.

Requirements

• macOS with Raycast
• A running LightRAG instance (see the upstream project for installation and configuration)
• Network access from your Mac to that server (e.g. http://127.0.0.1:<port> for local installs)

Setup

1. Open Raycast → Extensions → LightRAG Knowledge Base → Preferences (or the extension’s settings).
2. Set LightRAG Server URL to your API base URL, including protocol and port (example: http://127.0.0.1:9621 — use the port your server exposes).
3. Enter LightRAG Username and LightRAG Password as configured on your server.

Features

• LightRAG Status — Command to check connection and server health.
• Raycast AI tools — Natural-language queries, structured retrieval (entities/chunks/references), document listing, pipeline and track status, inserting text notes, and graph helpers (labels, subgraph, entity checks). Tool behavior is described in ai.yaml for the AI.

Privacy

Credentials are stored in Raycast preferences and never committed to this repository. Use only servers and accounts you trust.

Third‑party

LightRAG is an open-source project by third parties; this extension is not affiliated with or endorsed by the LightRAG authors. Use their documentation for server setup, models, and API behavior.

Development

npm install
npm run dev    # open in Raycast development mode (keeps tools compiled while running)
npm run build  # writes compiled JS to ~/.config/raycast/extensions/lightrag/
npm run lint

License

This project is licensed under the MIT License — see LICENSE.