Book Fetch MCP

Example Use Cases

• 1

  Fetch and read the full text of published books within Claude or other MCP clients.

• 2

  Handle very long books by caching and paginating content to avoid exceeding MCP context limits.

• 3

  Build a RAG-style QA workflow over books, including planned ingestion, chunking, and reranking capabilities.

Description

This MCP server lets agents access the contents of published books from within Claude or other MCP clients, caching long texts and supporting pagination for very large works. It is designed to enable question-answering over books, with planned features for using large-context models (e.g., Gemini 1.5) and an internal mini-RAG ingestion/chunking/reranking pipeline. The project also mentions future integration with Sci-Hub to broaden access to papers. Setup instructions show how to run the server using the 'uv' runner and a provided MCP command configuration.