mcp_docs_reader
If you are the rightful owner of mcp_docs_reader and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
A lightweight MCP server for loading, processing, and querying PDF documents for Claude Desktop.
The mcp-docs-reader is a Model Context Protocol server designed to facilitate document-based question answering by integrating with Claude Desktop. It processes PDF files from a local directory, extracts and chunks their content, and builds a semantic search index using FAISS. The server then retrieves relevant document passages based on user queries and constructs a prompt to send to Claude. This setup allows users to interact with their documents in a more meaningful way, leveraging the power of semantic search and natural language processing. The server is easy to set up, requiring minimal configuration, and utilizes the uv tool for environment management.
Features
- Loads and processes PDF documents from a local docs/ folder
- Extracts text and splits it into semantic chunks
- Generates vector embeddings using SentenceTransformer
- Builds a FAISS-based vector index for semantic search
- Retrieves top-k relevant chunks based on user query
