What gets sent where
Plainly, so you can decide for yourself.
Leaves your Mac
| Meeting audio | To OpenAI, to be transcribed and answered. This is the other side of the conversation |
|---|---|
| Your context | Your CV, the job description and panel details accompany each question as instructions |
| Session timings | To our server, to meter minutes — start, stop and duration. No content |
| Your email | To our server, to identify the account |
Never leaves your Mac
| Your microphone audio | Excluded from the upload by default. The toolbar shows Mic⊘ when it's excluded |
|---|---|
| Your own words | Transcribed by macOS on-device |
| Recordings and transcripts | Written to your Documents folder |
| Your CV file | The text is sent as context, but the file stays put and we never store it |
Changing the default. Settings ▸ Your voice ▸ Send to model can be
switched to include your microphone. The app states in plain words when your voice will be
uploaded — it's never a silent change.
What we hold
Your email, your minute balance and ledger, a token per device, and session start/stop times. Not your CV, not your questions, not your answers, not your recordings. Full detail in the privacy policy.
What OpenAI holds
They process the audio and context to produce answers. Under their API terms, data sent through the API isn't used to train their models. Our key never leaves our server — the app connects with a token that expires in about a minute.