Cursor Chat History

This MCP server provides AI assistants access to Cursor chat history for analysis and insights, enabling personalized coding assistance based on actual development patterns. Built using TypeScript with better-sqlite3 and the MCP SDK, it reads local Cursor conversation databases directly without external services, supporting both legacy and modern conversation formats with tools for listing, searching, and retrieving full conversations, plus advanced analytics including usage patterns, file activity, programming language distribution, and relationship detection between conversations. The implementation uses a summary-first approach for context efficiency, allowing users to browse conversation summaries before diving into full content, and includes data extraction capabilities with JSON, CSV, and graph export formats, making it valuable for generating personalized coding rules from actual usage patterns, extracting best practices from past problem-solving sessions, and creating project documentation based on real development conversations.