enterprise

This MCP server exposes two main tools for CrewAI workflows: kickoff_crew and get_crew_status. kickoff_crew starts a deployed CrewAI workflow, while get_crew_status polls and returns the status and results of a running crew. To integrate with Claude Desktop, configure the MCP server in Claude’s Developer Settings by pointing it to the server URL and providing the required bearer token. Once configured, Claude can call the two tools to kick off a crew or check its status and retrieve results. The server is designed to manage Crewai Enterprise workflows, enabling you to initiate deployments and inspect results directly from your MCP-enabled tooling.

In practice, you would typically provide the necessary input parameters (such as identifiers for the specific CrewAI workflow or deployment) to kickoff_crew, then periodically request status updates with get_crew_status until the results are returned or the workflow completes. Ensure the MCP URL and authentication token are kept secure and are refreshed as needed according to your security policies.

Prerequisites:

• Python 3.8+ (or a compatible Python runtime) and Git installed on your development machine
• Access to the CrewAI Enterprise MCP server repository (cloned locally)

1. Install MCP tooling locally

• Install the MCP core and CLI components used by this server:

  # Using pip (Python tooling)
  python -m pip install --upgrade pip
  python -m pip install mcp
  python -m pip install mcp[cli]
  

  or follow the repository’s preferred installation method if provided.

2. Clone the enterprise MCP server repository

git clone https://github.com/crewaiinc/enterprise-mcp-server.git
cd enterprise-mcp-server

3. Prepare environment variables

• Create a local environment file or export variables in your shell:

  export MCP_CREWAI_ENTERPRISE_SERVER_URL="https://your-crewai-enterprise-server.example.com"
  export MCP_CREWAI_ENTERPRISE_BEARER_TOKEN="your-token-here"
  

4. Run locally with the MCP runner (as described in the README)

# Ensure you have the repository path ready and replaced in the command below
uv run --with mcp[cli] mcp run <path-to-cloned-repo> /crewai_enterprise_server.py

5. Verify installation

• Use the MCP tooling to list servers and ensure the crew tools (kickoff_crew, get_crew_status) are available and responding.

Notes:

• Ensure network access to the CrewAI Enterprise server URL and that the bearer token is valid.
• If running in a container or CI, consider using a dedicated environment and secret management for the token.

Tips and common considerations:

• Required env vars: MCP_CREWAI_ENTERPRISE_SERVER_URL and MCP_CREWAI_ENTERPRISE_BEARER_TOKEN. Keep tokens secure and rotate per your security policy.
• The server exposes two tools: kickoff_crew (to start a CrewAI workflow) and get_crew_status (to fetch status and results).
• When configuring in Claude Desktop, ensure the path to the cloned repository and the server file (/crewai_enterprise_server.py) are correct.
• If you encounter authentication errors, verify the token scope and that the URL is reachable from your execution environment.
• For local development, ensure you have the MCP CLI installed and that your Python environment has access to the required dependencies for the CrewAI enterprise integration.