Every meeting assistant listens. This one answers.
Recording, transcription and a tidy summary are settled — every tool in the category does them, and does them well enough. What none of them do is speak up while the meeting is still happening. MeetCrew joins as a named, disclosed participant and answers when somebody addresses it, because behind it is the agent your team already built.
The notes were never the hard part.
Ask any meeting assistant a question during the call and nothing happens — it is a recorder. The answer arrives afterwards, in a document, once the moment that needed it has passed. That gap is the whole reason this product exists.
Answers in the meeting
Say its name and it responds out loud, while the question still matters. Then it goes quiet and carries on taking the notes.
Remembers across meetings
Decisions, action items and open questions carry from one call to the next — and your own agent can read that memory back over MCP and act on it.
Runs your model, not ours
Point it at your Copilot Studio, Microsoft Foundry, or custom agent. Your prompts, your knowledge, your data boundary.
Disclosed on camera, and no voiceprints.
A meeting assistant that quietly records is a problem waiting to be discovered. This one is visible by design: named in the participant list, badged as AI for the entire call, and it says out loud what it is when it joins. Anyone in the meeting can tell it to leave.
Speakers are linked by platform identity — never by anything derived from a voice. That is a commitment in the architecture, not a checkbox in an admin console.
All four on the same plan, with the same teammate and the same memory.
No client plugin and no per-user add-in. It joins from a link, or from a calendar invite.
Not per seat, and not per minute of recording.
See it next to Otter, Fireflies, Read.ai, Fathom and Zoom AI Companion.
Try it against the assistant you already use.
Bring both to the same meeting. The difference shows up the first time somebody asks it a question.