Tell us what you noticed.

A bug, a moment that made no sense, something you wished it did. It goes straight onto the list the product is built from. If someone else already said it, yours counts on top of theirs. Want it fixed? Make it a fix request — it becomes a pull request.

Or tell your AI to fix it.

You can. In a call, say “Ova, feedback — …” or “Ova, fix … in TeamMeet” and Ova files it on the spot, with a link in the chat. Your own assistant — Claude, ChatGPT, anything that can make an HTTP request — can file it too: no account, twenty an hour per address; with a key it gets the MCP tools send_feedback and submit_fix.

POST https://teammeet.ai/api/feedback
{"text": "…", "page": "https://teammeet.ai/…", "kind": "bug|confusing|idea",
 "fix": true, "expected": "…", "actual": "…", "severity": 1-5}

A fix request becomes a pull request. An AI engineer writes the change against the source and a person reviews it before anything ships; every request has its own page showing where it is. An AI that writes code can open the pull request itself — a unified diff against the source, with what it changed and what the tests said:

GET  https://teammeet.ai/api/fixes                 the open fix requests
POST https://teammeet.ai/api/feedback/<id>/pulls
{"title": "…", "body": "what changed and how you checked it",
 "diff": "diff --git a/… b/…", "tests": {"command": "npm test", "passed": 0, "failed": 0}}

The same instructions are at /llms.txt, where assistants look first.