Neura Market
modelcontextprotocol/server-sqlite logo

modelcontextprotocol/server-sqlite

Free

官方参考实现，操作 SQLite 数据库，并内置商业智能能力。

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About modelcontextprotocol/server-sqlite

This is an official reference implementation of a Model Context Protocol (MCP) server for SQLite, developed by the Model Context Protocol organization. It provides database interaction and business intelligence capabilities, allowing users to run SQL queries, analyze business data, and automatically generate business insight memos through a continuously updated memo resource. The server exposes a dynamic resource (memo://insights), a demonstration prompt (mcp-demo) for guided analysis, and six core tools: read_query, write_query, create_table, list_tables, describe-table, and append_insight. It can be run locally using uv or Docker and integrates seamlessly with MCP-compatible clients like Claude Desktop or VS Code. Note: This repository has been archived and is now read-only as of May 29, 2025.

Key Features

Runs SQL queries (SELECT, INSERT, UPDATE, DELETE, CREATE TABLE) directly through MCP tools
Exposes a dynamic memo resource (`memo://insights`) that aggregates business insights discovered during analysis, auto-updating as new insights are appended
Provides a demonstration prompt (`mcp-demo`) that guides users through database operations, schema creation, and business analysis for a given topic
Six core tools: read_query, write_query, create_table, list_tables, describe-table, and append_insight
Can be installed and run via `uv` (Python) or Docker with simple configuration in Claude Desktop or VS Code
Open source and free to use, with full source code available on GitHub

Pros & Cons

Pros
  • Official reference implementation, well‑documented and aligned with MCP standards
  • Simple setup with both `uv` and Docker options, suitable for various development environments
  • Built‑in business intelligence features that automatically compile insights into a memo resource
  • Includes a guided demo prompt for easy onboarding and exploration of database operations
  • Lightweight and self‑contained (uses SQLite), no external database server required
Cons
  • Repository is archived and no longer actively maintained (read‑only since May 2025)
  • Limited to SQLite databases, not suitable for production‑scale relational databases
  • Setup assumes users are familiar with MCP configuration (claude_desktop_config.json, Docker, uv)
  • Only provides basic SQL operations; advanced features like transactions, views, or indexes are not explicitly exposed as tools

Best For

Setting up a local SQLite database for experimentation or testing with MCP-enabled AI assistantsPerforming business data analysis by querying SQLite databases and automatically generating insight memosCreating and managing database schemas through natural language interactions in Claude Desktop or VS CodeEducational demonstrations of database concepts and SQL operations using the mcp-demo promptIntegrating SQLite database access into AI-driven tools and workflows via the Model Context Protocol

FAQ

What is the Model Context Protocol (MCP) server for SQLite?
It is an official reference MCP server that enables AI assistants (like Claude Desktop) to interact with SQLite databases, execute SQL queries, and generate business insights automatically.
How do I install the SQLite MCP server?
Two methods are provided: using `uv` with a Python setup, or using Docker. Configuration is added to your `claude_desktop_config.json` (for Claude Desktop) or VS Code settings.
What tools does the server provide?
It provides six tools: read_query (execute SELECT), write_query (INSERT/UPDATE/DELETE), create_table, list_tables, describe-table (schema info), and append_insight (add to memo).
Can I use the server with other MCP‑compatible clients besides Claude Desktop?
Yes, the server follows the MCP standard and can be used with any client that supports MCP, including VS Code as shown in the documentation.
Is this server still actively maintained?
No, the repository was archived by the owner on May 29, 2025, and is now read-only. It may still work but will not receive updates.