Model providers
Effective November 3, 2026
These are the companies Genien sends data to, what each one does, and where to read its rules. When you use a model, you also follow the usage policy of the company that makes and hosts it, as the Terms of Service say. The Privacy Policy explains what we send and why.
1. Providers that process data for Genien
| Provider | What it does in Genien | Rules | Privacy |
|---|---|---|---|
| Supabase | Accounts, database and file storage | Terms | Privacy policy |
| Render | Hosting the app | Terms | Privacy policy |
| Anthropic | Claude writes briefs, ideas, scripts and ad packs, and checks frames and words | Usage policy Commercial terms | Privacy policy |
| fal.ai | Runs the image, video, voice, music and clean-up models you pick on the canvas | Terms API services terms | Privacy policy |
| OpenRouter, with Google's Gemini | Watches and listens to generated videos for the quality check | OpenRouter terms Gemini API terms | Privacy policy |
| Creatomate | Assembles the final video, captions and end cards | Terms | Privacy policy |
| YouTube | Plays model makers' videos on our home page and in one guide | Terms | Privacy policy |
2. Model makers reached through fal.ai
fal.ai runs models from Kling, Google (Nano Banana, Gemini, Lyria), OpenAI (GPT Image, Whisper), Black Forest Labs (FLUX), ByteDance (Seedream, OmniHuman), Alibaba (Wan, Qwen), MiniMax, Luma, Ideogram, Recraft, xAI, Lightricks, Stability AI, Topaz Labs, Sync, VEED, Creatify, HeyGen, Bria, Clarity AI, Pixelcut, ElevenLabs, Mirelo, CassetteAI, Sonilo and open-source models. Each model's license and use rules are on its page at fal.ai, and two makers publish rules that apply to every use of their models:
- Google: Generative AI Prohibited Use Policy
- OpenAI: Usage policies
Upscaling and enhancement models from Topaz Labs: Powered by Topaz Labs.
3. Training and retention
We use these companies' API services, not their consumer apps. Where a provider lets us choose, we use settings that keep your data from being used to train its models. Each provider keeps data under its own retention schedule, described in the policies linked above. Models from providers whose terms we haven't cleared are on hold and aren't offered.
4. Changes
We add a provider here before it receives your data. The models page shows every model on the canvas, its maker and whether it's on hold.