### Install Dependencies with uv

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Installs project dependencies using the 'uv' package manager.

```bash
# Install dependencies using uv
uv sync
```

--------------------------------

### Run MCP Server Directly with uv

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Starts the MCP server directly using 'uv run' with the 'mcp[cli]' extra.

```bash
# Run the MCP server directly
uv run --with mcp[cli] mcp run crewai_enterprise_server.py
```

--------------------------------

### kickoff_crew

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Starts a new crew execution by sending a POST request to the /kickoff endpoint. It accepts task inputs and returns a crew_id for status tracking.

```APIDOC
## kickoff_crew

### Description
Sends a POST request to the `/kickoff` endpoint of the configured CrewAI Enterprise server to start a new crew execution. The `inputs` dictionary is forwarded as the task payload. The response includes a `crew_id` that must be saved and used to poll for the task's result via `get_crew_status`.

### Method
POST

### Endpoint
`/kickoff`

### Request Body
- **inputs** (dict) - Required - The payload for the crew task.

### Request Example
```json
{
  "inputs": {
    "query": "Summarize the latest advancements in quantum computing",
    "output_format": "bullet_points",
    "max_sources": 5
  }
}
```

### Response
#### Success Response (200)
- **crew_id** (string) - The unique identifier for the initiated crew task.
- **status** (string) - The current status of the crew task (e.g., "started").
- **message** (string) - A confirmation message.

#### Response Example
```json
{
  "crew_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "status": "started",
  "message": "Crew task has been initiated successfully"
}
```
```

--------------------------------

### Get Crew Task Status and Results

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Retrieves the current status and output of a crew task using its 'crew_id'. Sends a GET request to the server's /status/{crew_id} endpoint. Requires 'httpx' and 'python-dotenv'.

```python
import asyncio
import httpx
import os
import time
from dotenv import load_dotenv

load_dotenv()

CREWAI_ENTERPRISE_SERVER_URL = os.getenv("MCP_CREWAI_ENTERPRISE_SERVER_URL")
CREWAI_ENTERPRISE_BEARER_TOKEN = os.getenv("MCP_CREWAI_ENTERPRISE_BEARER_TOKEN")

async def get_crew_status(crew_id: str) -> dict:
    async with httpx.AsyncClient() as client:
        response = await client.get(
            f"{CREWAI_ENTERPRISE_SERVER_URL}/status/{crew_id}",
            headers={
                "Authorization": f"Bearer {CREWAI_ENTERPRISE_BEARER_TOKEN}",
                "Content-Type": "application/json",
            },
        )
        response.raise_for_status()
        return response.json()
```

--------------------------------

### get_crew_status

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Retrieves the status and results of a crew task using a GET request to the /status/{crew_id} endpoint. Requires the crew_id obtained from the kickoff_crew operation.

```APIDOC
## get_crew_status

### Description
Sends a GET request to `/status/{crew_id}` on the CrewAI Enterprise server to retrieve the current state of a running or completed crew task. The response contains the task status and, once complete, the full output produced by the crew's agents.

### Method
GET

### Endpoint
`/status/{crew_id}`

### Parameters
#### Path Parameters
- **crew_id** (string) - Required - The unique identifier of the crew task.

### Response
#### Success Response (200)
- **status** (string) - The current status of the crew task (e.g., "running", "completed", "failed").
- **output** (object or null) - The results of the crew task if completed, otherwise null.

#### Response Example
```json
{
  "status": "completed",
  "output": {
    "summary": "Advancements in quantum computing include breakthroughs in qubit stability and error correction techniques..."
  }
}
```
```

--------------------------------

### Run MCP Server as Python Script with uv

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Executes the MCP server as a Python script using 'uv run', initiating the MCP transport loop.

```bash
# Or run as a Python script (starts the MCP transport loop)
uv run crewai_enterprise_server.py
```

--------------------------------

### Kickoff Crew Task with Inputs

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Initiates a new crew execution by sending a POST request to the server's /kickoff endpoint. The 'inputs' dictionary is passed as the payload. Requires 'httpx' and 'python-dotenv'.

```python
import asyncio
import httpx
import os
from dotenv import load_dotenv

load_dotenv()

CREWAI_ENTERPRISE_SERVER_URL = os.getenv("MCP_CREWAI_ENTERPRISE_SERVER_URL")
CREWAI_ENTERPRISE_BEARER_TOKEN = os.getenv("MCP_CREWAI_ENTERPRISE_BEARER_TOKEN")

async def kickoff_crew(inputs: dict) -> dict:
    async with httpx.AsyncClient() as client:
        response = await client.post(
            f"{CREWAI_ENTERPRISE_SERVER_URL}/kickoff",
            headers={
                "Authorization": f"Bearer {CREWAI_ENTERPRISE_BEARER_TOKEN}",
                "Content-Type": "application/json",
            },
            json={"inputs": inputs},
        )
        response.raise_for_status()
        return response.json()

# Example: kick off a research crew with a topic query
async def main():
    result = await kickoff_crew({
        "query": "Summarize the latest advancements in quantum computing",
        "output_format": "bullet_points",
        "max_sources": 5
    })
    print(result)
    # Expected output:
    # {
    #   "crew_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
    #   "status": "started",
    #   "message": "Crew task has been initiated successfully"
    # }
    crew_id = result["crew_id"]
    print(f"Crew started with ID: {crew_id}")

asyncio.run(main())
```

--------------------------------

### Verify MCP Server Tools with mcp CLI

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Verifies that the MCP server exposes the expected tools by inspecting it with the 'mcp dev' command.

```bash
# Verify the server exposes the two tools by inspecting with mcp CLI
uv run --with mcp[cli] mcp dev crewai_enterprise_server.py
# Expected: lists kickoff_crew and get_crew_status as available tools
```

--------------------------------

### Register CrewAI Enterprise MCP Server in Claude Desktop

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Configure Claude Desktop to recognize and use the CrewAI Enterprise MCP Server. This involves specifying the command to run the server and its environment variables.

```json
// Claude Desktop: Settings > Developer Settings > MCP Servers
// claude_desktop_config.json
{
  "mcpServers": {
    "crewai_enterprise_server": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "/path/to/cloned/repo/crewai_enterprise_server.py"
      ],
      "env": {
        "MCP_CREWAI_ENTERPRISE_SERVER_URL": "https://your-crew-endpoint.app.crewai.com",
        "MCP_CREWAI_ENTERPRISE_BEARER_TOKEN": "your_bearer_token_here"
      }
    }
  }
}
```

--------------------------------

### Configure CrewAI Enterprise MCP Server Locally

Source: https://github.com/crewaiinc/enterprise-mcp-server/blob/main/README.md

Use this JSON configuration to set up the CrewAI Enterprise MCP Server when running locally. Ensure you replace placeholders with your actual file path and credentials.

```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": "<>"
      }
    }
  }
}
```

--------------------------------

### Configure CrewAI Enterprise Server Environment Variables

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

Set these environment variables to configure the connection to your CrewAI Enterprise backend. Ensure these are set before running the server.

```bash
# .env file (loaded automatically via python-dotenv)
MCP_CREWAI_ENTERPRISE_SERVER_URL=https://your-crew-endpoint.app.crewai.com
MCP_CREWAI_ENTERPRISE_BEARER_TOKEN=your_bearer_token_here
```

--------------------------------

### Poll Until Crew Task is Complete

Source: https://context7.com/crewaiinc/enterprise-mcp-server/llms.txt

This function polls for the status of a crew task, returning the result when completed or raising an error if failed. It uses an asyncio sleep to control the polling interval.

```python
async def wait_for_crew(crew_id: str, poll_interval: int = 5) -> dict:
    while True:
        status_response = await get_crew_status(crew_id)
        print(f"Current status: {status_response.get('status')}")

        if status_response.get("status") == "completed":
            print("Crew finished! Results:")
            print(status_response.get("result"))
            # Expected output:
            # {
            #   "crew_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
            #   "status": "completed",
            #   "result": "• Quantum error correction has advanced significantly...\n• New qubit designs...",
            #   "completed_at": "2024-11-15T10:32:00Z"
            # }
            return status_response
        elif status_response.get("status") == "failed":
            raise RuntimeError(f"Crew task failed: {status_response.get('error')}")

        await asyncio.sleep(poll_interval)

asyncio.run(wait_for_crew("a1b2c3d4-e5f6-7890-abcd-ef1234567890"))
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.