CrewAI Enterprise MCP Server
Official# CrewAI Enterprise MCP Server
## Overview
A Model Context Protocol (MCP) server implementation that provides deployed CrewAI workflows. This server enables kicking off your deployed crew and inspect the status giving the results of your crew.
## Tools
- kickoff_crew
- get_crew_status
## Env Variables
retrieve from app.crewai.com
`MCP_CREWAI_ENTERPRISE_SERVER_URL`
`MCP_CREWAI_ENTERPRISE_BEARER_TOKEN`
# Usage with Claude Desktop
To use this MCP server with Claude Desktop, follow these steps:
1. Open Claude Desktop
2. Go to Settings > Developer Settings
3. Add a new MCP server with the configuration shown below
## Locally, cloned repo:
Install `mcp` and `mcp[cli]` locally
```json
{
"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": "<>",
"MCP_CREWAI_ENTERPRISE_BEARER_TOKEN": "<>"
}
}
}
}
```
## TODO: Added on PyPI:
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: one starts a new crew task (kickoff_crew) and the other checks the status of an existing task (get_crew_status). There is no overlap or ambiguity between these operations.
Both tools follow a consistent verb_noun pattern (kickoff_crew, get_crew_status) with clear action-oriented verbs and the same noun root. The naming is predictable and follows the same convention throughout.
With only two tools, the server feels severely under-equipped for an 'Enterprise MCP Server' that presumably manages crew tasks. There are likely missing operations like listing crews, updating tasks, or handling errors that would be needed for complete workflow coverage.
The tool surface is significantly incomplete for crew task management. While it covers starting and checking status, there are obvious gaps: no way to list existing crews, update tasks, cancel tasks, or retrieve results beyond status. This will cause agent failures in many scenarios.