dbt MCP

The recommended approach is to run the server directly via npx using the published npm package. Alternatively, download the experimental dbt-mcp.mcpb bundle from the latest GitHub release assets.

npx -y @us-all/dbt-mcp

The server needs to know where your dbt project lives and how to connect to your warehouse. Make sure your dbt profiles.yml is configured correctly and that 'dbt debug' passes before adding the MCP server.

dbt debug  # Verify dbt connection works

Add the dbt MCP server to your Claude Desktop configuration file. On macOS this is at ~/Library/Application Support/Claude/claude_desktop_config.json. Point to your dbt project directory and include any required environment variables for dbt Cloud if needed.

{
  "mcpServers": {
    "dbt": {
      "command": "npx",
      "args": ["-y", "@us-all/dbt-mcp"],
      "env": {
        "DBT_PROJECT_DIR": "/path/to/your/dbt/project",
        "DBT_PROFILES_DIR": "~/.dbt"
      }
    }
  }
}

For Discovery API, Admin API, and Semantic Layer tools you need a dbt Cloud service token. Set the relevant environment variables so the server can authenticate with the dbt Platform APIs.

{
  "mcpServers": {
    "dbt": {
      "command": "npx",
      "args": ["-y", "@us-all/dbt-mcp"],
      "env": {
        "DBT_PROJECT_DIR": "/path/to/your/dbt/project",
        "DBT_PROFILES_DIR": "~/.dbt",
        "DBT_CLOUD_SERVICE_TOKEN": "your-dbt-cloud-service-token",
        "DBT_CLOUD_ACCOUNT_ID": "your-account-id"
      }
    }
  }
}

Fully quit and reopen Claude Desktop. The dbt MCP server exposes tools across SQL, Semantic Layer, Discovery, CLI, Admin, Codegen, LSP, and Docs categories. You can ask Claude to list available dbt models, run builds, or query metrics.