Trino MCP Server

Model Context Protocol server for Trino, providing AI models with structured access to Trino’s distributed SQL query engine.

⚠️ BETA RELEASE (v0.1.2) ⚠️
This project is stabilizing with core features working and tested. Feel free to fork and contribute!

Features

• ✅ Fixed Docker container API initialization issue! (reliable server initalization)
• ✅ Exposes Trino resources through MCP protocol
• ✅ Enables AI tools to query and analyze data in Trino
• ✅ Provides transport options (STDIO transport works reliably; SSE transport has issues)
• ✅ Fixed catalog handling for proper Trino query execution
• ✅ Both Docker container API and standalone Python API server options

Quick Start

# Start the server with docker-compose
docker-compose up -d

# Verify the API is working
curl -X POST "http://localhost:9097/api/query" \
     -H "Content-Type: application/json" \
     -d '{"query": "SELECT 1 AS test"}'

Need a non-containerized version? Run the standalone API:

# Run the standalone API server on port 8008
python llm_trino_api.py

LLM Integration

Want to give an LLM direct access to query your Trino instance? We’ve created simple tools for that!

Command-Line LLM Interface

The simplest way to let an LLM query Trino is through our command-line tool:

# Simple direct query (perfect for LLMs)
python llm_query_trino.py "SELECT * FROM memory.bullshit.real_bullshit_data LIMIT 5"

# Specify a different catalog or schema
python llm_query_trino.py "SELECT * FROM information_schema.tables" memory information_schema

REST API for LLMs

We offer two API options for integration with LLM applications:

1. Docker Container API (Port 9097)

The Docker container exposes a REST API on port 9097:

# Execute a query against the Docker container API
curl -X POST "http://localhost:9097/api/query" \
     -H "Content-Type: application/json" \
     -d '{"query": "SELECT 1 AS test"}'

2. Standalone Python API (Port 8008)

For more flexible deployments, run the standalone API server:

# Start the API server on port 8008
python llm_trino_api.py

This creates endpoints at:

• GET http://localhost:8008/ - API usage info
• POST http://localhost:8008/query - Execute SQL queries

You can then have your LLM make HTTP requests to this endpoint:

# Example code an LLM might generate
import requests

def query_trino(sql_query):
    response = requests.post(
        "http://localhost:8008/query",
        json={"query": sql_query}
    )
    return response.json()

# LLM-generated query
results = query_trino("SELECT job_title, AVG(salary) FROM memory.bullshit.real_bullshit_data GROUP BY job_title ORDER BY AVG(salary) DESC LIMIT 5")
print(results["formatted_results"])

This approach allows LLMs to focus on generating SQL, while our tools handle all the MCP protocol complexity!

Demo and Validation Scripts 🚀