SQLite MCP Server

SQLite MCP Server is a Model Context Protocol server, originally part of Anthropic's reference implementations, that gives AI assistants direct access to local SQLite databases. It lets AI agents run SELECT queries, execute INSERT/UPDATE/DELETE operations, create tables, list database schemas, and inspect table structures, all through natural language in tools like Claude Desktop, Cursor, and other MCP clients. The server provides a clean separation between read and write operations. The read_query tool handles SELECT and CTE (WITH) queries, while write_query covers data modifications. A dedicated create_table tool handles schema creation through standard SQL DDL statements. The list_tables and describe_table tools give AI agents full visibility into database structure, enabling context-aware query generation without manual schema documentation. SQLite MCP Server is Python-based, distributed through PyPI, and runs via uv (the Rust-based Python package manager) for fast, isolated execution. It requires only Python 3.7 or later and SQLite3 (bundled with Python), making it one of the simplest MCP servers to set up. Because SQLite databases are single files on your local filesystem, the server pairs well with Anthropic's filesystem MCP for complete file and database workflows. It is a strong choice for prototyping, local data analysis, personal knowledge bases, and any workflow where you want an AI agent to query structured data without standing up a remote database.