Run dbt MCP locally

Back to guides

dbt MCP

AI

Quickstart

Beginner

Menu

This quickstart walks you through connecting dbt MCP server to your local dbt project. This setup gives you dbt command tools (run, build, test, compile, and more) inside your AI assistant.

If you'd like to connect to dbt platform with the CLI, see the OAuth quickstart.

To connect to dbt Wizard, see the dbt Wizard quickstart.

No clone required

You don't need to clone the dbt-mcp repository. Install uv and run uvx dbt-mcp — it fetches and runs dbt-mcp for you.

Prerequisites

• Install uv
• A local dbt project (the folder containing your dbt_project.yml file)
• dbt installed and available locally

For the full list of environment variables and how to enable or disable toolsets, see the Environment variables reference.

Step 1: Find your paths

You need two values before configuring your MCP client:

• DBT_PROJECT_DIR — the full path to your dbt project folder (where dbt_project.yml lives). For example, if your project name is jaffle_shop, the path should be /Users/yourname/dbt-projects/jaffle_shop.
• DBT_PATH — the full path to your dbt executable.

macOS/Linux

# Find DBT_PATH
which dbt
# Example output: /opt/homebrew/bin/dbt

# Find DBT_PROJECT_DIR — run from inside your project folder
pwd
# Example output: /Users/yourname/projects/my_dbt_project

Report incorrect code

Windows

# Find DBT_PATH
where dbt
# Example output: C:\Python39\Scripts\dbt.exe

# Find DBT_PROJECT_DIR — run from inside your project folder
cd
# Example output: C:\Users\yourname\projects\my_dbt_project

Report incorrect code

Note: Use forward slashes in your configuration: C:/Python39/Scripts/dbt.exe

Step 2: Add to your MCP client

Replace the paths below with the values from Step 1:

Claude Desktop

1. In Claude Desktop, go to Settings → Developer tab → Edit Config.
2. Paste the following configuration, replacing the paths with your actual values:

{
  "mcpServers": {
    "dbt": {
      "command": "uvx",
      "args": ["dbt-mcp"],
      "env": {
        "DBT_PROJECT_DIR": "/path/to/your/dbt/project",
        "DBT_PATH": "/path/to/your/dbt/executable"
      }
    }
  }
}

Report incorrect code

3. Save and restart Claude Desktop.

Config file location:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Claude Code

Run this command, replacing the paths with your actual values:

claude mcp add dbt \
-e DBT_PROJECT_DIR=/path/to/your/dbt/project \
-e DBT_PATH=/path/to/your/dbt/executable \
-- uvx dbt-mcp

Report incorrect code

Cursor

Click the link below with Cursor open to auto-configure:

Add dbt v1 or dbt v2 to Cursor

After clicking:

1. Update DBT_PROJECT_DIR with the full path to your dbt project.
2. Update DBT_PATH with the full path to your dbt executable (from Step 1).
3. Save the configuration.

VS Code

1. Open Settings → Features → Chat and ensure MCP is enabled.
2. Open the Command Palette (Ctrl/Cmd + Shift + P) and select MCP: Open User Configuration.
3. Add the following configuration to mcp.json, replacing the paths with your actual values:

VS Code uses "servers", not "mcpServers"

{
  "servers": {
    "dbt": {
      "command": "uvx",
      "args": ["dbt-mcp"],
      "env": {
        "DBT_PROJECT_DIR": "/path/to/your/dbt/project",
        "DBT_PATH": "/path/to/your/dbt/executable"
      }
    }
  }
}

Report incorrect code

4. Save the file.

Step 3: Test your setup

Ask your AI assistant to run a dbt command (for example, "Run dbt compile on my project" or "List all models in my project"). If dbt MCP is working, the assistant will execute the command against your local project.

 Optional: verify from the command line

uvx dbt-mcp

Report incorrect code

If there are no errors, your configuration is correct. Press Ctrl+C to stop the server.

What's available

With CLI-only setup, your AI assistant can use:

• dbt run, dbt build, dbt test, dbt compile, dbt list, dbt parse, dbt show
• Model lineage and node details from your local project
• Codegen tools (when enabled — see Environment variables reference)

Platform features like Semantic Layer, Discovery API, and metadata queries require a dbt platform account. To add them, see Connect to dbt platform.

Troubleshooting

Can't find the uvx executable

Symptoms: Error messages like Could not connect to MCP server dbt-mcp, Error: spawn uvx ENOENT, or spawn uvx ENOENT in your MCP client.

Cause: Your MCP client (like Claude desktop) can't find uvx in its PATH because it starts with a limited environment.

Solution: Use the full path to uvx in your configuration.

1. Find the full path:

   • macOS/Linux: Run which uvx in Terminal.
   • Windows: Run where uvx in Command Prompt or PowerShell.

2. Replace "command": "uvx" with the full path:

{
  "mcpServers": {
    "dbt": {
      "command": "/full/path/to/uvx",
      "args": ["dbt-mcp"],
      "env": { }
    }
  }
}

Report incorrect code

Example on macOS with Homebrew: "command": "/opt/homebrew/bin/uvx"

For VS Code (mcp.json), the same fix applies — replace uvx with its full path in the command field.

Server not starting

Symptoms: The MCP server shows as disconnected or unavailable in your client.

Diagnosis: Check the server logs:

• VS Code: Open the Command Palette (Ctrl/Cmd + Shift + P) → MCP: List Servers → click the dbt server to see detailed logs.
• Claude Desktop: Check ~/Library/Logs/Claude (macOS) or %APPDATA%\Claude\logs (Windows).
• All clients: Set DBT_MCP_LOG_LEVEL=DEBUG in your environment variables to get more verbose output.

Common causes:

• Missing or incorrect (Applies to dbt v1.11 and later) DBT_ENGINE_PROJECT_DIR or DBT_PATH — verify the paths exist and are absolute paths.
• Invalid or expired authentication tokens — generate a new token and update your config.
• Missing required environment variables for the toolset you're trying to use — see Tool requirements.

For all troubleshooting topics, see dbt MCP FAQs and troubleshooting.

Next steps

• Add dbt platform features: see Connect to dbt platform
• Configure toolsets or disable specific tools: see the Environment variables reference

Was this page helpful?

Privacy policyCreate a GitHub issue

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.