Postgres MCP

1. A lightweight MCP (Model Context Protocol) server built to query Postgres database.
2. It provides tools for running raw SQL queries, introspecting schema, and querying using plain English.
3. You can use any OpenAI compatible LLM provider for English to SQL translation. This is fully automated, and you don't need to tell it anything.

Available Tools

Tool	Description
execute_raw_query	Run arbitrary SQL and return rows
list_tables	List all non-system tables
describe_table	Get column names, types, and nullability for a table
execute_english_query	Translate natural language to SQL and optionally execute

Minimum needed environment variables

1. POSTGRES_HOST
2. POSTGRES_USER
3. POSTGRES_PASSWORD
4. POSTGRES_DATABASE
5. LLM_API_KEY: Needed only if you want English language query capability
6. LLM_API_URL: Needed only if you are using any other LLM provider other than OpenAI

Other environment variables

1. POSTGRES_PORT
2. LLM_MODEL
3. POSTGRES_SSL (boolean)
4. POSTGRES_VERIFY (boolean)
5. POSTGRES_SEND_RECEIVE_TIMEOUT (in seconds)
6. POSTGRES_CONNECT_TIMEOUT (in seconds)

Configuring using mcpServers json

For Mac and Linux

{
  "mcpServers": {
    "postgres-mcp": {
      "command": "npx",
      "args": ["-y", "@freakynit/postgres-mcp@latest"],
      "env": {
        "POSTGRES_HOST": "<POSTGRES HOST>",
        "POSTGRES_USER": "<POSTGRES USER>",
        "POSTGRES_PASSWORD": "<POSTGRES PASSWORD>",
        "POSTGRES_DATABASE": "<POSTGRES DATABASE>",
        "LLM_API_KEY": "<LLM API KEY if using english language queries>",
        "LLM_API_URL": "<LLM API ENDPOINT if not using OpenAI>"
      }
    }
  }
}

For Windows

{
  "mcpServers": {
    "postgres-mcp": {
      "command": "cmd",
      "args": ["/k", "npx", "-y", "@freakynit/postgres-mcp@latest"],
      "env": {
        "POSTGRES_HOST": "<POSTGRES HOST>",
        "POSTGRES_USER": "<POSTGRES USER>",
        "POSTGRES_PASSWORD": "<POSTGRES PASSWORD>",
        "POSTGRES_DATABASE": "<POSTGRES DATABASE>",
        "LLM_API_KEY": "<LLM API KEY if using english language queries>",
        "LLM_API_URL": "<LLM API ENDPOINT if not using OpenAI>"
      }
    }
  }
}

---

Development Docs

Prerequisites

• Node.js
• A PostgreSQL server
• An OpenAI (or any OpenAI compatible provider) API key and the endpoint (if not using OpenAI)

Installation

git clone https://github.com/freakynit/postgres-mcp.git
cd postgres-mcp
npm install

Configuration

Create a .env file in the project root with the following variables:

# PostgreSQL
POSTGRES_HOST=your-db-host
POSTGRES_PORT=5432                      # Optional, default 5432
POSTGRES_USER=your-username
POSTGRES_PASSWORD=your-password
POSTGRES_DATABASE=your-database

# SSL settings (optional)
POSTGRES_SSL=false                      # Or true
POSTGRES_VERIFY=false                   # Verify server cert if SSL true

# Timeouts (optional, seconds)
POSTGRES_SEND_RECEIVE_TIMEOUT=60
POSTGRES_CONNECT_TIMEOUT=60

# OpenAI
LLM_API_KEY=sk-...                      # Needed only if you want English language query capability
LLM_API_URL=https://api.openai.com/v1   # Optional, needed only if you are using any other LLM provider other than OpenAI
LLM_MODEL=gpt-4o                        # Optional, default gpt-4o

Usage

Start the server (Starts in stdio mode):

node src/index.js

Project Structure

postgres-mcp/
├── src/
│   ├── config.js         # Environment parsing (Zod)
│   ├── db.js             # Postgres client factory
│   ├── openai_client.js   # OpenAI client factory
│   ├── register_tools.js  # Tool registration orchestrator
│   ├── index.js          # Entry point
│   └── tools/            # Individual tool definitions
│       ├── execute_raw_query.js
│       ├── list_tables.js
│       ├── describe_table.js
│       └── execute_english_query.js
├── .env.example          # Make sure to rename to `.env` and update it with correct values
├── package.json
└── README.md

Debugging with MCP inspector

npx @modelcontextprotocol/inspector node src/index.js

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/YourFeature)
3. Commit your changes (git commit -m "feat: ...")
4. Push to your branch (git push origin feature/YourFeature)
5. Open a Pull Request

License

This project is licensed under the MIT License. See LICENSE for details.