Marks, consent and your data
AI videos should be honest about what they are, and nobody’s face or voice should be used without their say. This is how ClonifyLab makes sure.
How our videos are marked
Content Credentials (C2PA)
Every file we make is signed with C2PA Content Credentials: metadata that says it was made by ClonifyLab and contains content generated by AI. Any C2PA tool can read it.
An invisible watermark
The picture carries a Video Seal watermark and the sound an AudioSeal watermark, with one public key. They are built to survive the re-encoding that social platforms do, and a free script finds them.
The platform’s AI label
When we post, each platform’s own AI label is switched on automatically wherever the law or the platform requires it. At present it goes on every Instagram and Facebook post, and on YouTube and TikTok on every post of a clone, a cloned voice or realistic AI visuals; only a stylized faceless video with a library voice goes without it there. Nobody can switch a required label off.
- We never add AI tags to your captions, titles or videos. You may say it in your own words.
- TikTok drafts can’t carry the label, so we remind you to switch on TikTok’s AI-generated content setting when you publish.
- If you post a file we delivered yourself, you switch on the platform’s AI label where it’s required and keep the marks in it, as our Terms say.
- On our own website, app and bot, every clone, cloned-voice or realistic example says “AI-generated with ClonifyLab” next to it.
How consent works
- A clone is made only from one continuous recording that starts with the person reading a consent sentence and a random four-word phrase shown on screen. We check the phrase in the transcript.
- Everyone clones only themselves. Business presenters record their own consent; a company can never consent for them.
- Everyone must be 18 or over. We never clone public figures, children, people who have died, or anyone without their own consent recording.
- To make a clone we ask for no ID document and do no face matching. Only when we have a specific reason do we ask an account to verify its identity with Stripe Identity, which checks an ID document and a selfie.
- Consent can be withdrawn at any time, in the app, the bot or by email. We stop at once and delete the clone and its videos within 30 days, at our providers too.
- We keep only the consent part of the recording, as proof, for 3 years after the clone is deleted.
- Clone mode is open to people who live in the EU.
- Read the Likeness & Voice Licence.
How approval works
- Nothing posts until its owner approves it, video by video and platform by platform.
- The preview shows the exact file, each platform’s text, the posting time and the platforms that get the AI label.
- We record who approved, when, and a fingerprint (hash) of the exact file and texts. An approved video can’t be changed.
- Businesses approve only through the people they list.
Where your data goes
We store data in the EU by default: our database in Ireland, media storage in the EU, our server in Germany and customer video renders in the Netherlands. Not all data stays in the EU: some of the AI providers we use (Anthropic, OpenAI, ElevenLabs, fal and Google’s Gemini API) process data in the US, under the EU’s Standard Contractual Clauses.
- The original clone recording is deleted 30 days after its clips are cut.
- Your clone and every video made with it are deleted within 30 days after your plan ends or you withdraw.
- Videos you never approve are deleted after 90 days.
- We don’t sell data and don’t share it with advertisers. Our website sets no cookies.
- All the details, providers and your rights: our Privacy Policy.
Safety checks and reports
- Every topic, script, post text and AI-made picture passes an automatic safety check: no sexual or political content, no public figures or other people’s faces, no fake testimonials, no income or unsupported medical or financial claims, no promised views. Your own clips and takes aren’t checked automatically: you confirm what they show when you upload them.
- When something is blocked, we say why and give the credit back. Anyone can ask a person to review a decision.
- Anyone can report a video or a misuse of their face or voice. We act within 48 hours.