MCP Trino Server

MCP Trino Server is an MCP server that connects AI assistants to distributed Trino SQL query engines and Apache Iceberg lakehouses. Designed for data engineers, analytics professionals, and developers building AI-assisted database workflows, it allows language models to interact directly with enterprise data infrastructures. Users can inspect schemas, explore catalogs, navigate metadata hierarchies, and execute analytical SQL queries with structured formatting. Beyond standard query operations, the server provides specialized utilities for Apache Iceberg table management. This includes inspecting internal table structures like snapshots, manifests, partitions, and file statistics, as well as initiating maintenance routines such as small file compaction via table optimization, manifest optimization, and snapshot expiration based on retention policies. The server supports multiple communication layers including standard input/output, streamable HTTP, and Server-Sent Events, making it adaptable to local IDE setups or distributed container environments running through Docker.

Category: Data & Analytics

Tags: big-data, data-lakehouse, iceberg, SQL, trino

Visit MCP Trino Server

How to install and configure MCP Trino Server

1. Clone the repository and configure your environment variables: bash git clone https://github.com/alaturqua/mcp-trino-python.git cd mcp-trino-python Create a .env file setting TRINO_HOST, TRINO_PORT, TRINO_USER, and optional default TRINO_CATALOG and TRINO_SCHEMA values. 2. To install automatically for Claude Desktop via Smithery, run: bash npx -y @smithery/cli install @alaturqua/mcp-trino-python --client claude 3. Alternatively, install dependencies locally using uv: bash uv sync 4. Add the server to your Claude Desktop configuration file: json { "mcpServers": { "trino": { "command": "python", "args": ["./src/server.py"], "env": { "TRINO_HOST": "your-trino-host", "TRINO_PORT": "8080", "TRINO_USER": "trino" } } } }

What you can do with MCP Trino Server

• Running ad-hoc SQL queries through Trino and reviewing formatted data analysis directly in an AI chat interface. - Exploring multi-tier database architectures by visualizing catalogs, schemas, tables, and views across your organization. - Compact small files and optimize Apache Iceberg manifests to maintain optimal lakehouse query performance. - Inspect Iceberg snapshots, partitions, metadata logs, and branch references to trace table lineage and changes over time. - Automate routine lakehouse data housekeeping by purging expired table snapshots based on configured age thresholds.

Key facts

• https://github.com/alaturqua/mcp-trino-python
• Data & Analytics, Databases & Data Stores
• big-data, data-lakehouse, iceberg, SQL, trino

Part of MCP Servers

How do I install MCP Trino Server?

You can install MCP Trino Server automatically for Claude Desktop using Smithery by running npx -y @smithery/cli install @alaturqua/mcp-trino-python --client claude. Alternatively, clone the GitHub repository and run it using uv sync or standard pip installations with Python 3.12 or higher.

What can MCP Trino Server do?

It provides tools to browse catalogs, schemas, and tables in Trino, execute arbitrary SQL queries, view table definitions, and inspect query history. It also supports deep Apache Iceberg integration, allowing users to view snapshots, data files, manifests, and partition details, while executing table optimization and snapshot expiration.

Which MCP clients work with MCP Trino Server?

It works with any Model Context Protocol compliant client. The documentation explicitly outlines configuration procedures for Claude Desktop and Visual Studio Code using standard input/output transport, as well as web clients or remote tools via streamable HTTP and Server-Sent Events transports.

Is MCP Trino Server open source?

Yes, MCP Trino Server is open source and licensed under the Apache 2.0 license. The source code, configuration files, and Docker deployment setups can be reviewed and contributed to directly on GitHub.