The CrewAI MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your CrewAI account. It provides structured and secure access to your deployed automations and deployment controls, so your agent can start executions, monitor progress, inspect deployment health, review logs, and resume paused work on your behalf.
- Automation execution: Have your agent check the required inputs and start a deployed CrewAI automation with the right information.
- Execution monitoring: Let the agent track an automation's state and progress, then retrieve its result or error.
- Deployment oversight: Instruct your agent to list available crew deployments and check whether a selected deployment is ready or experiencing problems.
- Deployment troubleshooting: Direct your agent to review deployment lifecycle logs when a build or runtime deployment fails.
- Recovery and continuation: Have your agent redeploy an existing automation when needed, or approve or retry a paused task so its execution can continue.