Every other notetaker on the market sends your audio to a transcription service and is billed by the audio-hour. That bill is why their free tiers have caps: each free meeting is a real invoice, so free users have to be rationed.
We do not have that bill. Whisper is downloaded once into your browser and runs on your own processor. Your audio never reaches us, so we never pay to transcribe it, so there is nothing for a weekly cap to recover. The free plan is uncapped for the boring reason that capping it would not save us anything.
The honest cost of that trade is on you rather than on us: a one-time model download before your first meeting, and slower transcription on an older machine than a datacentre GPU would give you. We would rather state that plainly here than discover it for you later.
What we do pay for, per call, is the hosted model behind summaries, action-item extraction, Ask and translation, and the servers behind cross-device sync and sharing. Pro unlocks that list, except AI summaries and action items, which we are covering on the free plan for now. Nothing in Pro is a feature we withheld from the free plan to create a reason to upgrade.