MCP Trino, developed by Tuannvm, is a free Model Context Protocol (MCP) server designed to integrate AI assistants with the Trino distributed SQL query engine. Built in Go, it allows AI to perform fast analytics on large datasets. This tool exposes essential database interaction capabilities through a suite of callable functions, supporting both STDIO and HTTP transport protocols. It is compatible with various AI clients, including Claude Code and Cursor, providing a bridge for intelligent data operations. MCP Trino offers optional OAuth 2.0 authentication, making it suitable for secure enterprise environments requiring identity management.

What does MCP Trino connect to?

MCP Trino functions as a Model Context Protocol server, establishing connections to Trino, formerly known as PrestoSQL. Trino is a powerful distributed SQL query engine specifically engineered for rapid analytical processing across extensive datasets. Its primary purpose within this setup is to empower AI assistants with direct access to Trino's query capabilities, enabling them to interact with and extract insights from large-scale data infrastructures. This integration allows AI systems to utilize advanced SQL analytics for various data-driven tasks.

Core features and tools

MCP Trino exposes several callable tools for AI interaction with Trino. These include execute_query (also known as execute_sql) for running SQL commands, list_catalogs, list_schemas, and list_tables for database introspection. Additionally, it offers get_table_schema (also known as describe_table) to retrieve detailed table structures, and explain_query for understanding query execution plans. It supports STDIO and Streamable HTTP endpoints at "/mcp", with an SSE endpoint at "/sse" for backward compatibility.

Who is it for?

MCP Trino targets AI assistants, AI engineers, and developers who require deep integration between AI capabilities and data infrastructure. It supports workflows where AI needs to query, analyze, and understand production data, or generate smart API responses based on database insights. Developers can also use it for intelligent schema design, allowing AI to examine existing structures and suggest new designs. This is not the right fit if you do not need AI to interact with a distributed SQL query engine.

How to get started

Access is granted through a standard OAuth authorization flow. Setup involves configuring environment variables like OAUTH_ENABLED, HTTPS_CERT_FILE, HTTPS_KEY_FILE, MCP_TRANSPORT, and MCP_URL for HTTP transport.

Use cases

• Interactive data exploration and analysis, allowing AI assistants to query, analyze, and understand production data.
• Smart API development, where AI can query databases to understand relationships, suggest JOIN strategies, and generate API responses.
• Intelligent schema design, enabling AI to examine existing table structures, identify conventions, and suggest new schema designs.

MCP Trino provides a clear path for AI systems to engage with Trino's powerful analytics. Its Go-based architecture and support for standard protocols like HTTP make it adaptable for various environments. The optional OAuth 2.0 authentication offers secure integration, ensuring that AI assistants can access data responsibly. With its ability to connect AI to large datasets, MCP Trino is a valuable tool for data-driven applications.