🚀 Trino MCP Server

A Model Context Protocol (MCP) server that gives AI assistants (like Claude in Cursor) direct access to query and explore Trino/Presto databases.

💡 No installation required! Just configure and run with uvx.

✨ Features

• 🔍 Schema Discovery - Explore schemas, tables, and columns
• 📊 Smart Exploration - Get schema + sample data in one call
• ⚡ Safe Query Testing - Automatic LIMIT for exploration queries
• 📈 Data Profiling - Null %, distinct counts, top values
• 💾 Export Results - Save to CSV, Excel, JSON, or Parquet
• 🔐 Encryption Check - Detect encrypted columns
• 📋 Query History - Find frequently used queries

🚀 Quick Start (3 minutes)

Prerequisites

• Cursor IDE or Claude Desktop
• Trino database credentials

Step 1: Install uv (One-time, 30 seconds)

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

💡 uv is a fast Python package manager. The uvx command lets you run Python tools without installing them permanently.

Step 2: Find your uvx path

which uvx
# Usually: ~/.local/bin/uvx (macOS/Linux) or %USERPROFILE%\.local\bin\uvx (Windows)

Step 3: Configure Cursor

Open your Cursor MCP settings:

macOS: ~/.cursor/mcp.json
Windows: %APPDATA%\Cursor\mcp.json

Add this config (replace the uvx path with YOUR path from Step 2):

{
  "mcpServers": {
    "trino": {
      "command": "/Users/YOUR_USERNAME/.local/bin/uvx",
      "args": ["trino-mcp-navi"],
      "env": {
        "TRINO_HOST": "your-trino-host.company.com",
        "TRINO_PORT": "443",
        "TRINO_USER": "your.email@company.com",
        "TRINO_CATALOG": "awsdatacatalog",
        "TRINO_DEFAULT_SCHEMA": "default"
      }
    }
  }
}

⚠️ Important: Use the full path to uvx (not just "uvx"). Cursor doesn't inherit your shell's PATH.

Step 4: Restart Cursor

Restart Cursor completely (close and reopen) to load the MCP server.

Step 5: Test Connection

In Cursor's AI chat, type:

Test my Trino connection

The AI will use mcp_trino_test_connection and confirm the connection.

That's it! 🎉 No pip install, no cloning repos.

---

🔧 Configuration Options

Environment Variable	Required	Default	Description
TRINO_HOST	✅	-	Trino server hostname
TRINO_PORT	✅	8080	Server port (usually 443 for HTTPS)
TRINO_USER	✅	-	Your username (usually email)
TRINO_PASSWORD	❌	-	Password (if using basic auth)
TRINO_CATALOG	✅	awsdatacatalog	Default catalog
TRINO_DEFAULT_SCHEMA	❌	default	Default schema
TRINO_HTTP_SCHEME	❌	https	http or https
TRINO_FALLBACK_HOST	❌	-	Backup host if primary fails

---

📋 Available Tools (18 Total)

🔍 Discovery Tools

Tool	Description
test_connection	Verify Trino connectivity
get_all_schemas	List all schemas in catalog
get_tables_in_schema	List tables in a schema
describe_table	Get column definitions
explore_table	Recommended: Schema + sample data in one call
search_catalog	Search for tables/columns by name

📊 Data Quality & Analytics

Tool	Description
profile_column	Null %, distinct values, top 5 values
profile_table	Row count, null analysis per column
get_table_stats	Partitions, distinct values, distribution
get_frequent_queries	Find how others query this table
check_encryption	Check for encrypted columns

⚡ Query Execution

Tool	Description
test_query	Quick test with auto-LIMIT (safe)
run_trino_query	Execute query, return results
save_query_results	Execute + export to CSV/Excel/JSON/Parquet
execute_sql_file	Run SQL from a file

🚀 Performance & Schema

Tool	Description
explain_query	Get execution plan
get_query_history	Recent queries from session
get_table_ddl	SHOW CREATE TABLE
compare_tables	Compare schemas of two tables

---

🎯 Example Usage

Explore a Table

Show me the schema and sample data from the users table in analytics schema

→ AI uses explore_table(table_name="users", schema_name="analytics")

Run a Query

Get the top 10 customers by revenue from last month

→ AI uses run_trino_query with appropriate SQL

Profile Data Quality

What's the null rate for email column in customers table?

→ AI uses profile_column(table_name="customers", column_name="email", schema_name="...")

Search for Tables

Find all tables with "order" in the name

→ AI uses search_catalog(search_term="order", object_type="table")

---

☁️ Cloud Hosting (Zero Install for Users)

Deploy the server to the cloud so users don't need to install anything!

Deploy to Railway (Recommended)

1. Click the button above or go to railway.app
2. Connect your GitHub repo
3. Set environment variables:
   • TRINO_HOST, TRINO_PORT, TRINO_USER, TRINO_CATALOG
4. Deploy! You'll get a URL like https://trino-mcp-navi.up.railway.app

Deploy to Render

1. Go to render.com
2. Create new Web Service → Connect GitHub repo
3. Use the render.yaml blueprint
4. Set environment variables
5. Deploy!

After Deployment - User Config

Users add this to their Cursor mcp.json - no installation needed!

{
  "mcpServers": {
    "trino": {
      "url": "https://your-deployment-url.railway.app/sse"
    }
  }
}

That's it! The server handles everything remotely.

---

🐍 Alternative: Run with Python

If you prefer not to use uvx:

# Install
pip install trino-mcp-navi

# Configure Cursor with python instead of uvx
{
  "mcpServers": {
    "trino": {
      "command": "python",
      "args": ["-m", "trino_mcp_navi"],
      "env": { ... }
    }
  }
}

---

🔧 Troubleshooting

"Connection refused" Error

• Check TRINO_HOST is correct
• Verify TRINO_PORT (usually 443 for HTTPS, 8080 for HTTP)
• Ensure network/VPN access to Trino server

"Authentication failed"

• Verify TRINO_USER and TRINO_PASSWORD
• Some Trino setups use SSO - check with your admin

MCP Not Loading in Cursor

1. Check you have uv installed: uv --version
2. Verify the JSON syntax in mcp.json is valid
3. Check Cursor logs: Help → Toggle Developer Tools → Console
4. Restart Cursor completely

"uvx: command not found"

Install uv first:

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

---

🔐 Security Notes

• Never commit credentials to version control
• Use environment variables for sensitive data
• The server connects via HTTPS by default
• All queries go directly to your Trino - no external services

---

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

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

---

📝 License

MIT License - see the LICENSE file for details.

---

🆘 Support

If you encounter issues:

1. Check the troubleshooting section above
2. Verify your Trino credentials work outside of Cursor
3. Test with a simple query: SELECT 1
4. Open an issue on GitHub

---

Made with ❤️ for data analysts and engineers