What we process, and why
To replace your head and hair with a persona in real time, virtualcam.ai analyses the geometry of your face in each video frame — position, pose, and expression — so the character can move and emote as you do. This processing exists for exactly one purpose: rendering your live character. It is never used to identify you, to recognise you across sessions or services, or to build a profile of you.
Consent
Facial processing happens only when you choose it: it begins when you start a live session and stops when the session ends. Do not start a session unless you agree to the processing described in this notice. We are adding a dedicated consent step - separate from accepting the Terms, and recorded with a timestamp - that will ask for your explicit agreement before your first live session. You can withdraw consent at any time by not using the camera features, or contact [email protected] to close your account. Never point virtualcam.ai at anyone else’s face without their clear, informed permission (see the Acceptable Use Policy): consent must come from the person whose face is processed.
Retention and destruction schedule
| Data | Lifetime |
|---|---|
| Facial geometry derived from a video frame | Exists in GPU memory only for the moments the frame is being transformed, then gone. Never written to storage, never persisted across sessions. |
| Your live video (source and transformed) | Processed in memory; not recorded, not stored. |
| Moderation-flagged still frames (if any) | Retained access-restricted for up to 90 days (or until an enforcement action concludes) for human review (see the Privacy Policy). These are images; we do not extract facial geometry from them. |
| Persona images (built-in today; your uploads, when that ships) | Uploaded persona images will be stored encrypted, tied to your account, and deleted when you delete them or your account (within 30 days). |
Our commitments
- We will never sell, lease, or trade biometric or facial data.
- We do not use your live video or facial data to train AI models without your explicit, separate opt-in consent.
- Facial processing happens on per-session cloud GPUs. Providers act only on our instructions to run your session; we do not permit them any independent use of your data.
- We do not use facial recognition: nothing in the service matches your face against a database or another person.
Questions and requests
For anything about facial data - including consent withdrawal and deletion requests - contact [email protected]. If someone has run your face through virtualcam.ai without your permission, report it here.