Search indexed documentation using natural language queries. Results are ranked by relevance and include source context. The limit parameter controls the number of results returned.
List all indexed documentation sources currently available in the system, providing information about each source.
Crawl and extract content from specified URLs, adding them to the processing queue. URLs must be valid and accessible.
Permanently remove documentation from the index based on URL. This action cannot be undone.
Display the current processing queue status, showing URLs waiting to be processed.
Process all URLs in the queue, extracting and indexing their content. Reports processing status and any errors encountered.
Clear all URLs from the processing queue. This permanently removes items from the queue, requiring re-adding if needed later.
RAG Documentation provides a powerful set of tools for retrieving and processing documentation through vector search. It enables AI assistants to augment their responses with relevant documentation context, improving accuracy and providing source-backed information. The server implements a Retrieval-Augmented Generation (RAG) system specifically designed for documentation management. It allows for crawling URLs to extract documentation, managing a processing queue, searching through indexed content with natural language queries, and maintaining the documentation database.
This MCP server provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context.
To use the RAG Documentation MCP server, you'll need to install it via npm:
npm install @hrudolph/mcp-ragdocs
After installation, you can configure your AI client to use this MCP server by adding the following configuration:
"mcpServers": {
"ragdocs": {
"command": "npx",
"args": ["@hrudolph/mcp-ragdocs"]
}
}
Once installed and configured, the RAG Documentation server provides several capabilities:
The server stores indexed documentation and vector embeddings locally. No additional configuration is required for basic usage.
For advanced usage, you may want to customize:
A typical workflow might look like:
extract_urlsrun_queuesearch_documentationlist_sourcesThis allows AI assistants to provide responses backed by specific documentation, improving accuracy and providing source attribution.