crystaldba/postgres-mcp

Overview

crystaldba/postgres-mcp is a Model Context Protocol server that provides configurable read and write access to PostgreSQL databases for AI agents. It also includes performance analysis capabilities, allowing agents to monitor and optimize query execution. Built in Python, it integrates with any MCP-compatible client.

Best for

Best for
Developers who want to give AI agents structured and safe access to PostgreSQL databases

Use cases

• Give an AI agent direct SQL query access to a Postgres database
• Automate database performance monitoring and slow query analysis
• Allow natural language database interaction through an MCP client like Claude

How to use

Install

uv pip install postgres-mcp

Tools exposed

• list_schemas
• list_objects
• get_object_details
• execute_sql
• explain_query
• get_top_queries
• analyze_workload_indexes
• analyze_query_indexes
• analyze_db_health

Tested with

Claude Desktop, Cursor, Windsurf, Cline, ChatGPT

Notes

crystaldba/postgres-mcp is a Model Context Protocol server that provides configurable read and write access to PostgreSQL databases for AI agents. It also includes performance analysis capabilities, allowing agents to monitor and optimize query execution. Built in Python, it integrates with any MCP-compatible client.

2,831 stars on GitHub. Last updated 2026-01-22. Licensed MIT.

Use cases

• Give an AI agent direct SQL query access to a Postgres database
• Automate database performance monitoring and slow query analysis
• Allow natural language database interaction through an MCP client like Claude

Pros

• Configurable read/write access levels give granular control
• Built-in performance analysis helps identify slow queries
• Open source with strong community adoption (over 2800 stars)

Cons

• Requires an MCP-compatible client to interact with the database
• Write access carries risk of accidental data modification or deletion
• Setup involves configuring a Python server and credentials

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