mcp-sqlite
MCP server that provides SQLite database operations. Allows AI assistants to query, modify and manage SQLite databases through the Model Context Protocol.
Installation
pip install -r requirements.txt
Quick Start
from mcpsqlite import SQLiteServer
server = SQLiteServer(db_path="data.db")
server.run()
Features
- Execute SELECT, INSERT, UPDATE, DELETE queries
- Schema inspection and table listing
- Transaction support with rollback
- Query result formatting
- Safe parameterized queries
MCP Tools
query
Execute a SQL query and return results.
{
"name": "query",
"arguments": {
"sql": "SELECT * FROM users WHERE age > ?",
"params": [18]
}
}
execute
Execute a SQL statement (INSERT, UPDATE, DELETE).
{
"name": "execute",
"arguments": {
"sql": "INSERT INTO users (name, age) VALUES (?, ?)",
"params": ["Alice", 25]
}
}
schema
Get database schema information.
{
"name": "schema",
"arguments": {
"table": "users"
}
}
tables
List all tables in the database.
{
"name": "tables",
"arguments": {}
}
Configuration
server = SQLiteServer(
db_path="data.db",
read_only=False,
max_rows=1000
)
Examples
Running as MCP server
python -m mcpsqlite --db data.db --port 8080
Claude Desktop configuration
{
"mcpServers": {
"sqlite": {
"command": "python",
"args": ["-m", "mcpsqlite", "--db", "data.db"]
}
}
}
License
MIT
