Flask error tracking your coding agent can read

The Runheed Python SDK hooks into Flask's own request-exception signal. Every unhandled exception is reported with the method and URL, and Flask still answers the request the way it always does: a 404 stays a 404 and is not reported. Your agent reads the errors over MCP.

Set it up

1.Install the SDK.

terminalbash

pip install https://runheed.com/dl/runheed-0.1.0-py3-none-any.whl

2.Initialize it and attach it to your app.

app.pypython

import os
import runheed
from flask import Flask
from runheed import FlaskIntegration

runheed.init(dsn=os.environ["RUNHEED_DSN"], environment="production")

app = Flask(__name__)
integration = FlaskIntegration()
integration.setup(runheed.get_client())
integration.init_app(app)

3.Connect your agent.

terminalbash

claude mcp add runheed --env RUNHEED_API_KEY=al_... -- npx -y https://runheed.com/dl/runheed-mcp-0.1.0.tgz

The npm and PyPI releases are not out yet, so these commands install the same package files straight from runheed.com. Nothing else changes when the registry versions ship.

What it catches

Unhandled exceptions in routes, with method and URL
Errors logged through the logging module (on by default)
Errors you report with runheed.capture_exception(e)

What it does not do

abort() and other HTTP errors are not reported, since they are responses you chose
Background jobs need capture_exception in their own except blocks
No performance tracing: Runheed tracks errors only

Then ask your agent

> what broke in production since the last deploy? fix the top one and resolve it with the commit hash

Common Flask errors

Plans from $29/mo Other agents and editorsCancel any time from the billing portal