dbt-mcp

Why we included this project

Data teams running dbt who want AI assistants to work inside their analytics stack are the audience here. The server exposes a dbt project to agents through tools for SQL execution, the Semantic Layer, and the Discovery API. An agent can list metrics, fetch compiled SQL, or turn natural language into queries, all grounded in the project's real context. Because dbt-labs maintains it, the server tracks the platform's actual APIs and works across dbt Core, Fusion, and dbt Platform. Teams already invested in dbt can hook it into their MCP-capable clients and give agents a dependable view of their data models and metrics.