mcp-monitor-sdk

Hosted monitoring SDK for Python MCP servers. User guide · Developer guide

Repo copies: USER.md · DEV.md

User guide

Create a server and API key at https://mcp-moniter.vercel.app/. Wrap your FastMCP server with this package. Tool-call metrics (name, duration, success/error) are batched and sent to the hosted API. They are not application logs. Open the same dashboard to see metrics and analytics for that key’s server.

pip install mcp-monitor-sdk

from mcp_monitor_sdk import MonitoredFastMCP

server = MonitoredFastMCP(
    "todo-mcp",
    api_key="replace-with-a-64-character-or-longer-api-key",
)

@server.tool()
def todos_list() -> str:
    """List all todos."""
    return "No todos."

if __name__ == "__main__":
    server.run()

Omit metrics_server_url in production. Default ingest is https://mcp-metrics-server.just-a-dev.workers.dev/v1/metrics.

Developer guide

Same payload as the JS SDK. Built to sit next to the official mcp package.

When developing this repo against a local API:

server = MonitoredFastMCP(
    "todo-mcp",
    api_key="replace-with-a-64-character-or-longer-api-key",
    metrics_server_url="http://localhost:8000/v1/metrics",
)

Options

Failed flushes are restored to the buffer. Process exit drains remaining events over a sync HTTP POST. Body is a JSON array of events.

Local checkout:

pip install -e ./cloud/sdk/py

Publish:

cd cloud/sdk/py
python -m build
twine upload dist/*