Echo is the cinematic transcription and meeting intelligence platform — every speaker attributed, every decision pulled, every moment searchable.
Each panel reads from the same active recording engine. The motion is the live signal.
Echo is opinionated about scope. Four jobs in the recording surface — the rest stays human.
Echo transcribes 24 languages with 47ms latency. The caption is on-screen before the speaker finishes the sentence — no dropped names, no half-quotes.
Echo separates voices on the first frame, learns names from your calendar invite, and carries identity across recordings. Every quote attributable; every meeting searchable.
Echo extracts decisions, owners, and dates into a clean recap. The 47-minute call becomes a six-line summary your team will actually read by lunch.
Type a phrase. Echo plays the exact six seconds it was said, by whom, and links back to the meeting. The org's verbal history becomes a real index.
We benchmarked the same call on Echo and on a typical Otter + manual notes stack.
"We searched our verbal archive for a quote we'd been arguing about for six weeks. Echo found it in two seconds, played the eight-second clip. The argument ended."
You pay for the recording owner. Guests are never billed; team members listen and search at no extra charge.
Echo runs an on-device model first (Whisper-3 distilled) for the live caption and a hosted model for the corrected transcript that arrives seconds later. The user sees the caption immediately; the archive gets the polished version.
Live captions are computed on the participant's device — audio never leaves the call. Archived transcripts are processed in your chosen region (EU, US, SG, AU). Studio customers can pin processing to their own VPC.
Echo opens with a clear participant-consent banner that respects EU two-party consent rules. Refusing consent stops Echo for that participant only — the meeting continues for everyone else.
Yes. Studio customers can route long-form transcription through their own self-hosted Whisper, NVIDIA Parakeet, or commercial ASR — Echo provides the diarization, decision-pull, and search layer on top.
Per seat per month, hours unlimited on Reel and Studio. The 'seat' is the recording owner — guests on a meeting are never billed.