CrewAI Enterprise MCP Server

How to use

To use this MCP server with your MCP client, connect via a local stdio-based runtime configuration that points to the server implementation. You will typically kick off a crew, then poll or fetch status updates to see results. Use the provided tools to manage workflow executions and check outcomes from your client.

How to install

Prerequisites - Python is installed on your system (for the server runtime script mentioned in the configuration). - You have an MCP runtime client available to connect to the local or remote MCP server. - Access credentials (URL and token) for your CrewAI Enterprise server are available to you. Step 1: Prepare the local stdio MCP configuration. Create or edit your MCP configuration to include the following server entry (adjust paths if you clone the repository locally):

{
  "mcpServers": {
    "crewai_enterprise": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "<filepath of cloned repo>",
        "/crewai_enterprise_server.py"
      ],
      "env": {
        "MCP_CREWAI_ENTERPRISE_SERVER_URL": "<>",
        "MCP_CREWAI_ENTERPRISE_BEARER_TOKEN": "<>"
      }
    }
  }
}

Additional setup notes

Placeholders in the environment block should be replaced with your actual server URL and bearer token. The <filepath of cloned repo> should point to the location of the server script in your local clone. When you start the MCP client, it will interface with the server using the provided runtime configuration, allowing you to initiate and monitor CrewAI workflows.

Notes on security and usage

Treat the Bearer Token like a password. Do not expose the URL or token in client-side code or public repositories. Use secure storage for your credentials and rotate tokens regularly according to your organization’s security policies.

Troubleshooting tips

If you encounter authentication errors, verify that MCP_CREWAI_ENTERPRISE_SERVER_URL and MCP_CREWAI_ENTERPRISE_BEARER_TOKEN are correct and that the server script path is valid. If the server process fails to start, check that the uv runtime is installed and accessible in your environment. Ensure the file path inside the command matches the actual location of the server script in your cloned repository.

kickoff_crew

Starts a deployed CrewAI workflow (crew) and returns an execution handle or identifier for status tracking.

get_crew_status

Retrieves the current status and results of a previously kicked-off CrewAI workflow.