Privacy and data
Your voice is analyzed on this device and never stored. Here is exactly what goes where, and everything you can control.
What leaves this device
- Always on this device: pace estimate, pauses, pitch, volume and your recordings.
- Only if your browser supports it: on-device speech recognition for fillers and hedge words.
What exactly is sent
Chrome sends your audio to Google, and Edge to Microsoft, to get words back. It gives the best filler-word detection, but your voice leaves this device.
What exactly is sent
After each take, the words you said and that take’s delivery numbers go to our AI coach (Anthropic’s Claude), which judges the answer the way the listener would and coaches what to say.
If your browser can’t turn speech into words on the device (Safari, iPhone, Firefox), the finished answer’s audio goes to our transcription provider, Deepgram, to get the words, and is deleted once transcribed.
Nothing is stored by SpeakWings, and no provider trains on it.
Voices
Two studio voices: your coach, and the interviewer or manager who asks the questions. Your personal feedback is voiced live; only that text is sent to our voice provider, and nothing is stored.
What exactly is sent
Event names and counts, like “finished a session”, and only if the site has a collection endpoint set. Never audio, words, scores or your name. When this is off, counts stay on this device and only appear in a pilot report if you paste one.
Voice time capsule
Optional. Keep your Session 1, 7, 14 and 21 recordings so you can hear the difference. They live only in this browser and are never uploaded.
Your data
Your progress is stored on this device, and in your account only if you sign in. You can export or erase everything at any time.