hannesrudolph/mcp-ragdocs

Free

提供通过向量搜索检索和处理文档工具的 MCP 服务器实现，使 AI 助手能用相关文档上下文增强响应。

FreeFree tier

Type

Open Source

Visit Website GitHub (40 stars)

About hannesrudolph/mcp-ragdocs

mcp-ragdocs is 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. It features vector-based documentation search and retrieval, support for multiple documentation sources, semantic search capabilities, automated documentation processing, and real-time context augmentation for LLMs. The server includes tools such as search_documentation, list_sources, extract_urls, remove_documentation, list_queue, run_queue, and clear_queue, allowing users to search indexed documentation, manage sources, extract URLs from web pages, and control a processing queue for batch indexing.

Key Features

Vector-based documentation search and retrieval

Semantic search capabilities

Support for multiple documentation sources

Automated documentation processing and indexing

Real-time context augmentation for LLMs

Queue management for batch URL processing

Tools: search, list sources, extract URLs, remove docs, manage queue

Pros & Cons

Pros

Provides semantic search for documentation with relevance ranking
Open-source and free to use
Easy to set up as MCP server for integration with AI systems
Includes queue management for batch processing of URLs
Supports multiple documentation sources

Cons

Requires running the MCP server infrastructure
Documentation sources need to be indexed before searching
May require additional setup for embedding models and vector storage
Limited to text-based documentation; does not process images or video

Best For

Enhancing AI responses with relevant documentation contextBuilding retrieval-augmented generation (RAG) applications for documentationIntegrating documentation search into AI assistants and chatbotsAutomated crawling and indexing of web documentation sources

FAQ

How does search_documentation work?

It searches through stored documentation using natural language queries, specific terms, or code snippets, and returns matching excerpts with context ranked by relevance. You can limit the number of results (1-20, default: 5).

What is the extract_urls tool used for?

It crawls a specified web page, identifies all hyperlinks, and optionally adds them to the processing queue for later indexing. Use with caution on large sites to avoid excessive queuing.

How does the processing queue work?

The queue stores URLs to be processed and indexed. You can list pending URLs with list_queue, process them with run_queue (sequential processing with error handling and retry), or clear the queue with clear_queue.

Can I remove indexed documentation?

Yes, use the remove_documentation tool with an array of exact URLs to permanently remove those sources from the system.