CrewAI MCP

How to Install and Use CrewAI MCP

If you want to manage your deployed CrewAI workflows easily, the CrewAI MCP (Model Context Protocol) server is a helpful tool. It lets you start your Crew deployments and check their status quickly. Here's how to install and use the CrewAI MCP with clear, step-by-step instructions and code examples.

Step 1: Prepare Your Environment

Before installing CrewAI MCP, you need to set up some environment variables. These variables allow the MCP server to connect securely to your CrewAI account.

You must get these from your CrewAI app at app.crewai.com. The two important variables are:

• MCP_CREWAI_ENTERPRISE_SERVER_URL
• MCP_CREWAI_ENTERPRISE_BEARER_TOKEN

Set these environment variables on your machine by running commands like this (replace <your_value> with the actual values):

export MCP_CREWAI_ENTERPRISE_SERVER_URL="<your_server_url>"
export MCP_CREWAI_ENTERPRISE_BEARER_TOKEN="<your_bearer_token>"

Step 2: Clone the Repository and Install MCP

Next, clone the CrewAI MCP server repository to your local computer. Open your terminal and run:

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

After cloning, install mcp and its command-line tools locally using this Python command:

pip install mcp mcp[cli]

Step 3: Run the CrewAI MCP Server

Now, you are ready to start the MCP server. You can do this using the command below. Replace <filepath of cloned repo> with the path where you cloned the repository:

uv run --with mcp[cli] mcp run <filepath of cloned repo> /crewai_enterprise_server.py

For example, if you cloned it in your Documents folder under enterprise-mcp-server, the command looks like this:

uv run --with mcp[cli] mcp run ~/Documents/enterprise-mcp-server /crewai_enterprise_server.py

Running this command starts the server that talks to your Crew deployments.

Step 4: Use CrewAI MCP with Claude Desktop

To get the most out of CrewAI MCP, you can connect it with Claude Desktop for easy access. Follow these steps:

1. Open Claude Desktop on your computer.
2. Go to Settings, then select Developer Settings.
3. Add a new MCP server by entering the following configuration:

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

Make sure to replace the placeholders with your actual file path and environment variable values.

Step 5: Kickoff Crew and Check Status

Once everything is set up, you can use the two main tools that CrewAI MCP offers:

• kickoff_crew: To start your Crew deployments.
• get_crew_status: To check the status and get results from your Crew.

This makes it easy to manage your AI workflows directly from the MCP server.

---

By following these easy steps, you will have CrewAI MCP installed and ready to use. It's a powerful tool that helps you control and monitor your CrewAI deployments smoothly.