sqlite-mcp

An MCP (Model Context Protocol) server that exposes SQLite databases as tools for AI agents. Connect any SQLite database and let LLMs query schemas, run SQL, and inspect tables through a standardized interface.

• `execute_sql` — Run arbitrary SQL statements with support for dry-run mode (EXPLAIN QUERY PLAN), automatic result truncation, and configurable row limits.
• `get_table_info` — Retrieve complete table metadata in a single call: columns, types, DDL, indexes, foreign keys, and row count.
• Read-only mode — Optionally block all write operations for safe, production-friendly access.
• Query timeout — Configurable per-query timeout to prevent runaway queries.

• Python 3.11+
• uv (recommended) or pip

# Clone the repository
git clone https://github.com/<owner>/sqlite-mcp.git
cd sqlite-mcp

# Install with uv
uv sync

Or with pip:

pip install .

| Variable | Required | Default | Description | |---|---|---|---| | SQLITE_DB_PATH | Yes | — | Path to the SQLite database file | | SQLITE_READ_ONLY | No | false | Set to true to disable write operations | | SQLITE_TIMEOUT | No | 30 | Query timeout in seconds |

SQLITE_DB_PATH=./my_database.db sqlite-mcp

The server communicates over stdio and is designed to be launched by an MCP-compatible client (e.g., Cursor, Claude Desktop).

Add the server to your MCP client config (e.g., .cursor/mcp.json):

{
  "mcpServers": {
    "sqlite": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/sqlite-mcp", "sqlite-mcp"],
      "env": {
        "SQLITE_DB_PATH": "/path/to/database.db",
        "SQLITE_READ_ONLY": "true"
      }
    }
  }
}

Execute a SQL statement against the connected database.

| Parameter | Type | Default | D