Appearance
OpenRouter
OpenRouter (/media/openrouter) is a unified gateway to image models from many providers through a single OpenRouter API key. The studio exposes one text-to-image tab and discovers available image models live.
Where to configure
Configure your OpenRouter API key in Settings → AI. The media studio reuses the same OpenRouter LLM credential (universal-credential reuse), so configuring it once enables both AI chat and image generation. See Settings for provider setup.
Tabs / operations
| Tab | Operation | Key fields | Output |
|---|---|---|---|
| Text → Image | Image | Model (searchable combobox populated from OpenRouter's live catalog), prompt, number of images, aspect ratio, resolution, output format, optional seed | Still image(s) |
Generation flow
Image generation is synchronous. The finished image lands in /files via the Render Queue, where you can Edit in Designer or Post. See Media Studios overview for the shared flow.
Caveats
- Image only — OpenRouter does not expose video or audio generation through this studio.
- The model picker is populated live from
GET /api/v1/models, filtered to models that advertise image output. If the live catalog fails or a model is missing, the picker falls back to a static list and still allows you to type a model id directly. - Generation uses OpenRouter's dedicated
POST /api/v1/imagesendpoint, which returnsb64_jsondata. Each request waits up to 120 seconds (slow models such asgpt-image-1often need over a minute); a timeout or upstream failure is reported as an OpenRouter error naming the cause. - Because the credential is shared with the LLM provider, the studio is marked configured as soon as the OpenRouter AI key is present.
- The studio itself is text-to-image only. When OpenRouter is your workspace's image-to-image model default, AI Designer can condition generation on a reference image — supported for the Gemini image models (e.g.
google/gemini-2.5-flash-image), sent through OpenRouter's chat-completions endpoint; other models fall back to text-to-image.
Related docs
- Media Studios overview — shared render-queue and hand-off flow.
- Settings — configuring AI and media providers.
Verified against v1.3.0