mahAnuj/mcp-multi-db

Overview

A read-only MCP server that lets AI agents query PostgreSQL, MySQL, and SQLite databases. It enforces safety by default, preventing write operations while providing structured access to multiple database types through a single interface.

Best for

Best for
Developers who need a safe, multi-database read-only bridge for AI agents

Use cases

• Let an AI agent inspect production database schemas without risk of data modification
• Query multiple database types (PostgreSQL, MySQL, SQLite) from a single MCP endpoint
• Enable read-only data retrieval for AI-powered analytics or reporting tools

How to use

Install

npx -y mcp-multi-db

Tools exposed

• list_databases
• list_tables
• describe_table
• run_query

Tested with

Claude Desktop, Claude Code, Cursor, Cline, Windsurf

Example client config

{\n  "databases": [\n    {\n      "id": "local-sqlite",\n      "type": "sqlite",\n      "path": "/absolute/path/to/app.db",\n      "label": "Local dev SQLite"\n    },\n    {\n      "id": "analytics-pg",\n      "type": "postgres",\n      "connectionString": "postgresql://user:pass@localhost:5432/analytics",\n      "label": "Analytics Warehouse"\n    },\n    {\n      "id": "reporting-mysql",\n      "type": "mysql",\n      "connectionString": "mysql://user:pass@localhost:3306/reporting",\n      "label": "Reporting MySQL"\n    }\n  ]\n}

Notes

A read-only MCP server that lets AI agents query PostgreSQL, MySQL, and SQLite databases. It enforces safety by default, preventing write operations while providing structured access to multiple database types through a single interface.

0 stars on GitHub. Last updated 2026-06-21. Licensed ISC.

Use cases

• Let an AI agent inspect production database schemas without risk of data modification
• Query multiple database types (PostgreSQL, MySQL, SQLite) from a single MCP endpoint
• Enable read-only data retrieval for AI-powered analytics or reporting tools

Pros

• Read-only by design, reducing risk of accidental data corruption
• Supports three major database types in one server
• Lightweight TypeScript implementation with no external dependencies beyond database drivers

Cons

• No write support limits use cases to data retrieval only
• Zero stars and no community adoption suggests early or untested code
• Requires separate database credentials and network access for each database type

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.