dbt Docs MCP

Example Use Cases

• 1

  Search dbt nodes, column names, or compiled SQL across a dbt project to locate models, sources, and tests.

• 2

  Trace model- or column-level lineage (upstream predecessors and downstream successors) to understand dependencies and impact.

• 3

  Retrieve detailed attributes and metadata for a specific dbt node (unique_id), including schema mapping and catalog information.

Description

This server loads dbt Docs artifacts and exposes APIs to search dbt nodes, columns, and compiled SQL, retrieve node attributes, and explore model- and column-level lineage (upstream/downstream). It supports building column-level lineage from manifest and catalog files and integrates with MCP-capable clients (e.g., Claude desktop, Cursor) via a simple runner configuration. The tool is useful for automating metadata queries, tracing lineage, and surfacing schema and dependency information to agents and tooling.