PostgreSQL

What is PostgreSQL?

A Model Context Protocol server that provides read-only access to PostgreSQL databases. It enables LLMs to inspect database schemas and execute read-only SQL queries.

How to use PostgreSQL?

Configure the server in the mcpServers section of your claude_desktop_config.json using either Docker or NPX. Provide a PostgreSQL connection URL (e.g., postgresql://localhost/mydb). The server exposes a query tool for read-only SQL execution and automatically discovers table schemas as resources.

Key features of PostgreSQL

• Execute read-only SQL queries against the database
• Automatically discover table schemas (column names, data types)
• All queries run within a READ ONLY transaction
• Schema information exposed as JSON resources
• Supports Docker and NPX deployment methods

Use cases of PostgreSQL

• Let an LLM explore and query a PostgreSQL database safely
• Inspect database schema without write access
• Integrate read-only database queries into AI workflows

FAQ from PostgreSQL

What operations does the server allow?

The server only permits read-only SQL queries executed within a READ ONLY transaction. No write operations are possible.

How do I connect to a PostgreSQL database?

Use a PostgreSQL connection URL such as postgresql://user:password@host:port/db-name. For Docker on macOS, use host.docker.internal if the database runs on the host.

What schema information is exposed?

Table schemas are provided as JSON resources at postgres://<host>/<table>/schema, including column names and data types. Schemas are automatically discovered from database metadata.

What are the runtime requirements?

The server can be run via Docker (using the mcp/postgres image) or via NPX (using @modelcontextprotocol/server-postgres). No additional dependencies are listed beyond a PostgreSQL connection.

Is the server licensed?

Yes, it is licensed under the MIT License.