Client & sales calls
Contract terms, budgets and direct lines summarised without the client's name ever reaching a model provider's logs.
Content-blind meeting recorder You talk.
You talk.
The booth keeps it sealed.
The client call, the performance check-in, the clinical case conference — the meetings that most need a summary are exactly the ones you can't paste into a public AI notetaker. cloak-meet transcribes your audio entirely on your device, then cloaks every name, date and identifier with a realistic surrogate before anything is sent. The AI summarises a meeting between stand-ins; your device re-inks the real names back.
- 0 audio bytes posted
- 0 raw-PII bytes posted
- Signed on-device receipt
or
Nothing is uploaded. Recording and transcription both run in this tab.
Five sentences — what the meeting was and what came out of it.
No personal data detected — nothing to re-ink.
- on this device
- 0
- to the summariser
- 0
- anywhere else
- 0
Your summary will appear here.
The AI works from surrogates only, then your device re-inks the real names.
The meetings that can't go anywhere else
Performance check-ins
Honest one-to-ones written up as clean minutes, with the employee re-inked to a stand-in throughout.
Clinical case conferences
Patient discussions transcribed and summarised with DOB, refs and contact details cloaked from the first word.
Board & leadership updates
Decisions and action items pulled out in any of the languages the on-device model already speaks — Latin, Cyrillic, CJK and more.
From voice to summary, in four moves
Record — in the booth
MediaRecorder captures your microphone (or you upload an audio file) straight into browser memory. It is never uploaded, streamed or posted anywhere. Zero network.
Transcribe — on-device
A Whisper speech model (transformers.js, same-origin, no HF hub) turns the audio into text entirely in your tab. The audio never leaves — only the words it heard, staying on your device.
Cloak — on-device
A WebAssembly engine plus a multilingual name model find every name, email, phone, address and reference number in the transcript — across Latin, Cyrillic, Arabic and CJK — and write a realistic surrogate over each. Still zero network.
Summarise — re-inked
Only the surrogate transcript is sent, through the CloakAPI gateway with a cryptographic pre-tokenisation proof. The summary returns still carrying the surrogates; cloak-meet maps them back to your real values on-device and signs a receipt attesting 0 audio bytes and 0 raw-PII ever left.