An MCP server that provides tools for retrieving and processing documentation through vector search, both locally or hosted. Enabling AI assistants to augment their responses with relevant documentation context.
An MCP server implementation that provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context.
The RAG Documentation tool is designed for:
{
"mcpServers": {
"rag-docs": {
"command": "npx",
"args": ["-y", "@sanderkooger/mcp-server-ragdocs"],
"env": {
"EMBEDDINGS_PROVIDER": "ollama",
"QDRANT_URL": "your-qdrant-url",
"QDRANT_API_KEY": "your-qdrant-key" # if applicable
}
}
}
}
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"rag-docs-openai": {
"command": "npx",
"args": ["-y", "@sanderkooger/mcp-server-ragdocs"],
"env": {
"EMBEDDINGS_PROVIDER": "openai",
"OPENAI_API_KEY": "your-openai-key-here",
"QDRANT_URL": "your-qdrant-url",
"QDRANT_API_KEY": "your-qdrant-key"
}
}
}
}
{
"mcpServers": {
"rag-docs-ollama": {
"command": "npx",
"args": ["-y", "@sanderkooger/mcp-server-ragdocs"],
"env": {
"EMBEDDINGS_PROVIDER": "ollama",
"OLLAMA_BASE_URL": "http://localhost:11434",
"QDRANT_URL": "your-qdrant-url",
"QDRANT_API_KEY": "your-qdrant-key"
}
}
}
}
"ragdocs-mcp": {
"command": "node",
"args": [
"/home/sander/code/mcp-server-ragdocs/build/index.js"
],
"env": {
"QDRANT_URL": "http://127.0.0.1:6333",
"EMBEDDINGS_PROVIDER": "ollama",
"OLLAMA_URL": "http://localhost:11434"
},
"alwaysAllow": [
"run_queue",
"list_queue",
"list_sources",
"search_documentation",
"clear_queue",
"remove_documentation",
"extract_urls"
],
"timeout": 3600
}
| Variable | Required For | Default | remarks |
|---|---|---|---|
EMBEDDINGS_PROVIDER |
All | ollama |
“openai” or “ollama” |
OPENAI_API_KEY |
OpenAI | - | Obtain from OpenAI dashboard |
OLLAMA_BASE_URL |
Ollama | http://localhost:11434 |
Local Ollama server URL |
QDRANT_URL |
All | http://localhost:6333 |
Qdrant endpoint URL |
QDRANT_API_KEY |
Cloud Qdrant | - | From Qdrant Cloud console |
PLAYWRIGHT_WS_ENDPOINT |
Playwright Remote | - | WebSocket endpoint for remote Playwright server (e.g., ws://localhost:3000/) |
The repository includes Docker Compose configuration for local development:
docker compose up -d
This starts:
Access endpoints:
For production deployments:
QDRANT_URL=your-cloud-cluster-url
QDRANT_API_KEY=your-cloud-api-key
This project supports running Playwright either locally or via a Docker container. This provides flexibility for environments where Playwright’s dependencies might be challenging to install directly.
The src/api-client.ts file automatically detects the presence of the PLAYWRIGHT_WS_ENDPOINT environment variable:
PLAYWRIGHT_WS_ENDPOINT is set: The application will attempt to connect to a remote Playwright server at the specified WebSocket endpoint using chromium.connect(). This is ideal for using a containerized Playwright instance.PLAYWRIGHT_WS_ENDPOINT is not set: The application will launch a local Playwright browser instance using chromium.launch().A playwright service has been added to the docker-compose.yml file to facilitate running Playwright in a Docker container.
To start the Playwright server in Docker:
docker-compose up playwright
This command will pull the mcr.microsoft.com/playwright:v1.53.0-noble image and start a Playwright server accessible on port 3000 of your host machine.
To configure your application to use this containerized Playwright instance, set the following environment variable:
PLAYWRIGHT_WS_ENDPOINT=ws://localhost:3000/
Search through stored documentation using natural language queries. Returns matching excerpts with context, ranked by relevance.
Inputs:
query (string): The text to search for in the documentation. Can be a natural language query, specific terms, or code snippets.limit (number, optional): Maximum number of results to return (1-20, default: 5). Higher limits provide more comprehensive results but may take longer to process.List all documentation sources currently stored in the system. Returns a comprehensive list of all indexed documentation including source URLs, titles, and last update times. Use this to understand what documentation is available for searching or to verify if specific sources have been indexed.
Extract and analyze all URLs from a given web page. This tool crawls the specified webpage, identifies all hyperlinks, and optionally adds them to the processing queue.
Inputs:
url (string): The complete URL of the webpage to analyze (must include protocol, e.g., https://). The page must be publicly accessible.add_to_queue (boolean, optional): If true, automatically add extracted URLs to the processing queue for later indexing. Use with caution on large sites to avoid excessive queuing.Remove specific documentation sources from the system by their URLs. The removal is permanent and will affect future search results.
Inputs:
urls (string[]): Array of URLs to remove from the database. Each URL must exactly match the URL used when the documentation was added.List all URLs currently waiting in the documentation processing queue. Shows pending documentation sources that will be processed when run_queue is called. Use this to monitor queue status, verify URLs were added correctly, or check processing backlog.
Process and index all URLs currently in the documentation queue. Each URL is processed sequentially, with proper error handling and retry logic. Progress updates are provided as processing occurs. Long-running operations will process until the queue is empty or an unrecoverable error occurs.
Remove all pending URLs from the documentation processing queue. Use this to reset the queue when you want to start fresh, remove unwanted URLs, or cancel pending processing. This operation is immediate and permanent - URLs will need to be re-added if you want to process them later.
The package follows a modular architecture with clear separation between core components and MCP protocol handlers. See ARCHITECTURE.md{:target=“_blank”} for detailed structural documentation and design decisions.
curl -fsSL https://ollama.com/install.sh | sh
ollama pull nomic-embed-text
ollama list
This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.
We welcome contributions! Please see our CONTRIBUTING.md{:target=“_blank”} for detailed guidelines, but here are the basics:
npm installgit checkout -b feat/your-featureThis project is based on a fork of hannesrudolph/mcp-ragdocs, which itself was forked from the original work by qpd-v/mcp-ragdocs. The original project provided the foundation for this implementation.
A simple note-taking MCP server for recording and managing notes with AI models.
A open-source library enabling AI models to control hardware devices via serial communication using the MCP protocol. Initial support for Raspberry Pi Pico.
An MCP server that provides LLMs access to other LLMs
