POST /v1/images/generations endpoint is compatible with the OpenAI Images API. You change the model and pick from our catalog.
Which models can I use?
| ID | Provider | Price | Notes |
|---|---|---|---|
openai/gpt-image-1 | OpenAI | $0.19/img | Newest, high quality |
openai/dall-e-3 | OpenAI | $0.04/img | Classic |
google/imagen-4 | $0.04/img | Excellent with text | |
xai/grok-2-image | xAI | $0.07/img | Photorealistic style |
fal/flux-1.1-pro | fal.ai | $0.04/img | Best open-weight |
fal/flux-1.1-pro-ultra | fal.ai | $0.06/img | Flux 4K |
fal/recraft-v3 | fal.ai | $0.04/img | Best for typography and brand style |
fal/ideogram-v2 | fal.ai | $0.08/img | Professional typography |
fal/sd-3.5-large | fal.ai | $0.065/img | Stable Diffusion 3.5 |
Built-in storage
Each generated image is automatically uploaded to your Supabase Storage (supabase.geekhub.mx) in a public UUID-based bucket — effectively unguessable URLs (same model as Cloudinary or fal.media).
This means URLs don’t expire and you can reference the image from:
- Your web app (
<img src={url} />) - Slack, email, social networks
- Your own backup by downloading the bytes
Next steps
POST /v1/images/generations
Full endpoint reference.