Veymu
Create

Synthetic Media & Deepfake Policy

What we do to stop Veymu being used to fake real people, how a video traces back to the account that made it, and our response plan when one is reported.

Last updated October 10, 2026 · Operated by Ron Avidor (Israel)

The short version. Veymu makes AI video. Using it to make a real person appear to say or do something they didn't is against our Acceptable Use Policy. Every render is recorded against the account that made it, downloaded clips carry a note saying they came from Veymu, and a report to abuse@veymu.io is read by a person.

1. What we do to prevent misuse

  • Prompt screening. Every prompt is checked on our own servers before anything is rendered. A prompt that fails is refused and not charged.
  • The models' own safeguards. The video models we run refuse a reference picture or clip that appears to show a real person, unless the model's provider has verified that the picture is safe to use. They also refuse a wide range of other content independently of us.
  • Consent for uploads. Anyone who uploads a picture or clip of an identifiable person confirms that the person has consented to their likeness being used this way. Uploading someone else without consent breaks section 4 of the Acceptable Use Policy.

2. How a Veymu video traces back

We keep a record of each render in the account's library: the account, the prompt, the pictures or clips it used, the model and the time. When a finished render is deleted, from the library or with the whole account, a basic record of it (the account, the prompt, the model and the time) stays for 12 months, so a video made and then deleted can still be traced.

Clips downloaded from Veymu, and films exported from the Veymu editor, carry a short note in the file's metadata: “Made with Veymu (veymu.io). AI-generated video.”, followed by an identifier for the render. It doesn't change the picture or the sound, and it holds no personal details. If you report a file that still carries it, we can find the render it came from.

The note is ordinary metadata, not a visible watermark, and it is not a guarantee. Re-encoding a file, or uploading it to most social networks, removes it. A file without the note may still have come from Veymu, so report it anyway and we will search our records.

3. Reporting a deepfake

Email abuse@veymu.io with:

  • a link to the video, or the file itself;
  • where you came across it;
  • whether you are the person shown, or acting for them. Put that in the subject line: those reports go to the front of the queue, and we don't ask for proof of identity before taking content down.

4. What happens next

  1. We acknowledge the report within 2 business days. Reports involving a minor, or intimate imagery of a real person made without consent, are actioned the same day.
  2. We find the render from the note in the file, the link, or a search of our records.
  3. We review it against the Acceptable Use Policy.
  4. If it breaks the policy, we take it out of the account's library and stop serving it, and we suspend or close the account as section 5 of that policy describes. Content involving a minor or non-consensual intimate imagery closes the account permanently.
  5. We keep what the law needs. Where the law requires it, and always for content involving a minor, we preserve the record and report it to the relevant authorities. We cooperate with lawful requests from law enforcement and from the providers whose models we run.
  6. We tell you the outcome, as far as we can without disclosing another person's personal data.

5. Publishing AI video

Many countries and most large platforms now require realistic AI-made video to be labelled as such. Labelling where you publish is your responsibility, see section 7 of the Acceptable Use Policy. We recommend saying a video is AI-generated whenever a viewer might otherwise take it for a real recording.