Recording & consent
Recording, and the people in the room.
Astrix Live records your meetings so you can improve. Before capture begins, Astrix asks you to confirm that everyone has been informed, plays a voice notice to the room, or both.
What Astrix Live does
Astrix Live does not let anyone monitor another person. It scores you and only you. Before capture begins, Astrix asks you to confirm that everyone has been informed, plays a voice notice to the room, or both, and every other participant appears as an unnamed speaker label, because Astrix never collects their identity, never scores them and never builds a profile of them. Your audio is processed by the named sub-processors below, none of whom train on it.
Consent, and the law where you are
Recording other people carries obligations that vary by country and, in the United States, by state. Twelve states require every participant to consent: California, Connecticut, Delaware, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Oregon, Pennsylvania and Washington.
The in-app notice is your starting point, not the end of it. Say it in your own words too:
“I use a tool that scores how I ran this meeting. It records the call. Is that alright with everyone?”
Before recording, Astrix can ask you to confirm that everyone in the call has been informed and consents, play a voice notice to the room before capture begins, or both.
Where your audio goes
Astrix records two tracks on your computer: your microphone, and the call audio as it plays through your speakers. Both are sent over an encrypted connection to Astrix and transcribed by Soniox, our speech-to-text partner, whose real-time API does not retain audio or transcripts and does not train on them. Transcript text is analysed by the OpenAI API, which does not train on it but keeps abuse-monitoring logs for up to 30 days, and your microphone audio is analysed by our own voice-analysis model. The recordings, transcript and score are stored on Astrix infrastructure and saved to your account so you can track progress and link into practice.
Everyone else in the call appears as Speaker 1, Speaker 2 and so on. Astrix never asks who they are, never scores them and never builds a profile of them.
Sub-processors
Astrix uses the following sub-processors to run Astrix Live. We will give notice before adding a new one.
| Sub-processor | Purpose | What it receives |
|---|---|---|
| Soniox | Speech-to-text | Meeting audio: your microphone track and the call audio played on your computer. Real-time API: does not retain audio or transcripts, does not train on them. |
| OpenAI API | Transcript analysis | Transcript text only. Not used for training; abuse-monitoring logs kept for up to 30 days. |
| RunPod | GPU hosting for Astrix's own voice-analysis model | Your microphone audio and derived features, for scoring. |
| Cloudflare | Backend infrastructure and storage | Hosts the Astrix backend and stores recordings, transcripts and scores. |
Your data, and deleting it
For how Astrix protects and handles your data, see the Security & Trust page. To ask a question about your data or request deletion, contact us.