AI transparency
How vambel uses AI
vambel helps you make explainer videos with the help of AI. This page explains which AI providers are involved, what you must disclose when you publish, and how we protect real people. It follows the transparency rules of the EU AI Act (Article 50).
Which features use which providers
| Feature | Provider | Whose contract |
|---|---|---|
| In-app assistant, reading brand manuals, drafting a project from a website | Anthropic (Claude) | Yours: with your own Anthropic API key |
| Speech transcription of your recordings | Whisper, running on our own server | Ours: nothing leaves our server |
| Voice-overs, images, music, lip-synced presenter shots and video clips | ElevenLabs, OpenAI, xAI (Grok), BytePlus (Seedance), OpenRouter or Anthropic (Claude), with your own API key, depending on the feature and model you choose | Yours |
When you use your own API key, we transmit your content to that provider on your instruction, under your own contract with it. The provider bills you directly for usage.
No training on your content
We do not train AI models on your content. What providers you connect do with content is governed by your own contract with them.
Human oversight: nothing is published automatically
vambel never publishes anything on your behalf. You review every script, voice, image and clip, and you decide what to render and where to publish. AI output can be inaccurate or unexpected, so check facts, names and claims before you publish.
Real people and consent
- The face, likeness or voice of a real person may only be used with that person's documented consent.
- vambel records that consent with the personage and refuses lip sync and voice cloning without it.
- Deceptive deepfakes are not allowed: never make it look as if a real person said or did something they did not.
- No generated patient testimonials and no misleading health claims.
How disclosure appears in your films
Where generated people appear in a film, vambel adds an AI disclosure line to the end card.
What you must disclose when publishing
Under the EU AI Act, generated or manipulated video or audio that resembles real people, places or events, and could falsely appear authentic to viewers, must be clearly disclosed as AI-generated or manipulated. As the publisher, you remain responsible for this. In practice:
- keep the end-card disclosure, and do not remove it or cut the film in a way that hides it;
- where viewers might not see the end card (for example short clips or cut-downs), add a clear label such as "Made with AI" in the video itself or the post that accompanies it;
- use the AI-content labels of the platform you publish on, where available;
- for artistic, satirical or fictional work, disclose in a way that does not spoil the work but still makes clear AI was used.
Questions
Write to hello@vambel.com. To report misuse, write to abuse@vambel.com.