#949 · Primary category: AI Agents & Automation

pgmcp

agent agentic-ai ai analytics artificial-intelligence data-analysis database kong mcp mcp-server postgres postgresql

An MCP server to query any Postgres database in natural language.

Project last updated:05/26/26

GitHub Stars

540

Forks

62

Contributors

2

License

Apache-2.0

Why we included this project

PGMCP is an MCP server that lets AI assistants query an existing PostgreSQL database in plain English. If your team already works with Cursor, Claude Desktop, or VS Code extensions, you can point it at your current schema and ask questions like 'who are the top customers?' without writing SQL or changing the database. Results come back as tables, JSON, or CSV. The read-only enforcement and input validation are the details that matter if you are nervous about letting an agent near production data, and the schema cache plus auto-pagination keep large result sets manageable. It also handles free-text search across all columns and mixed-case table names, which is a common annoyance with real-world schemas. Teams that want a self-hosted data access layer for their AI tools, rather than a full BI dashboard, will find this a practical fit.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category