dbt MCP

Name: dbt MCP
Author: dbt Labs

by dbt Labs

Connect AI agents to your dbt data pipelines for metadata, lineage, and model exploration

data Python Intermediate Self-hostable Verified

⭐ 400 stars 📅 Updated: 1mo ago

View on GitHub View package

Description

Official dbt MCP server from dbt Labs that brings your data build tool environment directly into AI assistants. Access dbt metadata, explore data lineage graphs, query model definitions, review documentation, and understand transformation logic — all through natural language. Connects to dbt Cloud to provide rich context about your data models, their upstream and downstream dependencies, column-level lineage, and freshness status. Ideal for data engineers and analytics engineers who want AI assistance in navigating complex data warehouses and understanding transformation pipelines without manually browsing the dbt Cloud interface.

✅ Best for

Data and analytics engineers using dbt Cloud who want AI-assisted pipeline exploration

⏭️ Skip if

You use dbt Core without dbt Cloud, or your data stack does not include dbt

💡 Use cases

Exploring dbt model definitions and documentation via AI conversation
Tracing data lineage from source to final model through dependency graphs
Querying model metadata to understand column types, tests, and freshness
Getting AI summaries of complex transformation logic in dbt models

👍 Pros

✓ Official dbt Labs project with direct dbt Cloud integration
✓ Rich lineage and metadata access for comprehensive data understanding
✓ Helps onboard new team members to complex data pipelines via natural language
✓ Works with any dbt Cloud project regardless of warehouse backend

👎 Cons

✗ Requires dbt Cloud — does not work with dbt Core-only setups
✗ API token needs appropriate dbt Cloud permissions configured
✗ Large projects with thousands of models may have slow metadata responses

🔧 Exposed tools (5 tools)

Tool	Category	Description
list_models	discovery	List all dbt models in the project with summary metadata
get_documentation	documentation	Retrieve dbt documentation and descriptions for models and columns
explore_lineage	lineage	Explore upstream and downstream dependencies for a model
get_model_details	metadata	Get detailed information about a specific dbt model including columns and tests
query_model	query	Query model data and preview results from the data warehouse

⚡ Installation

Prerequisites:

• python v3.10+
• dbt Cloud credentials (API token and account ID)
• API key required

Check Claude Code documentation to configure this MCP server.

💡 Tips & tricks

Use a service token with metadata-only permissions for security. Start by asking the AI to map lineage for a specific model to understand dependencies quickly.

🔄 Alternatives

databricks-sdk-mcp

MCP server — editorial content pending

Similar functionality

bigquery-mcp