PostgreSQL

What is PostgreSQL?

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

How to use PostgreSQL?

Configure the server in the claude_desktop_config.json under mcpServers using either Docker (run mcp/postgres image) or NPX (@modelcontextprotocol/server-postgres). Provide a PostgreSQL connection URL (e.g., postgresql://host.docker.internal:5432/mydb) as an argument.

Key features of PostgreSQL

• Execute read-only SQL queries via the query tool.
• Automatically discover and serve table schemas as JSON resources.
• All queries run within a READ ONLY transaction.
• Supports Docker and NPX deployment methods.
• MIT licensed.

Use cases of PostgreSQL

• Let an LLM explore table structures and column types before generating queries.
• Allow an LLM to run safe, read-only SELECT statements against a live database.
• Provide database schema context to an AI assistant for answering questions about data.

FAQ from PostgreSQL

Is this server read-only?

Yes. All queries are executed inside a READ ONLY transaction, preventing any data modification.

What are the runtime requirements?

You need a running PostgreSQL database and either Docker or Node.js (for NPX). The server connects to the database via the provided URL.

How do I specify authentication credentials?

Include username and password in the PostgreSQL connection URL, e.g., postgresql://user:password@host:port/db-name.

Where does schema data come from?

Schema information is automatically discovered from the connected database's metadata. Each table's JSON schema includes column names and data types.

What transport does this server use?

The server follows the Model Context Protocol (stdio-based transport as typical for Claude Desktop MCP servers). The README does not detail other transport options.