MCP Server: PostgreSQL (and Aurora / Redshift)

What This Repository Does

@modelcontextprotocol/server-postgres is an official MCP server that connects build-cli to a PostgreSQL database (and compatible databases: Amazon Aurora PostgreSQL-compatible, Redshift with limitations), enabling direct SQL execution and schema exploration in a session.

Common use cases:

Ask questions about your data without writing SQL manually
Validate assumptions (“how many nulls are in this column?”) during development
Generate SQL against the actual schema rather than describing it in text
Explore an unfamiliar database

For the Snowflake-specific MCP server, see the companion item repo:mcp-snowflake.

How to Use It

Add the server to your Claude Code MCP configuration (~/.claude/settings.json):

{
  "mcpServers": {
    "postgres": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-postgres", "postgresql://user:password@host:5432/dbname"]
    }
  }
}

Replace the connection string with your credentials. Use a read-only database user where possible.

Security considerations:

Use read-only database credentials — MCP servers can execute arbitrary SQL.
Do not connect to production databases containing PII or regulated data without confirming compliance with your data steward and security team.
Credentials embedded in the config file are stored in plaintext — use environment variables or a secrets manager where possible.
Confirm with your team lead before connecting any MCP server to a production system.

Relevance to the Hub

This server enables DMDQP software engineers to query their PostgreSQL-family databases directly from build-cli, closing the loop between code authoring and data validation. It is especially useful during data quality investigations, schema-driven SQL generation, and ETL pipeline development.