Appearance
Fireworks AI
Fireworks AI (/media/fireworks) runs high-throughput open image models — including FLUX and Stable Diffusion XL — on Fireworks' inference-optimized infrastructure using your org's existing Fireworks AI key.
Where to configure
Configure the Fireworks AI provider once under Settings → AI. Fireworks is a universal-credential provider, so the same API key serves both LLM and media generation; no separate Settings → Media entry is required. See Settings for provider setup.
Tabs / operations
| Tab | Operation | Key fields | Produces |
|---|---|---|---|
| Text → Image | Image | Model (FLUX schnell/dev FP8, SDXL), prompt, aspect ratio, steps, guidance scale, seed | PNG image(s) in /files |
Generation flow
The studio calls Fireworks' workflow endpoint with Accept: application/json and receives base64-encoded images, which are decoded and saved to the Media Library. The Render Queue shows the result with Edit in Designer and Post options. See Media Studios overview for the shared flow.
Caveats
- Image only. Fireworks does not expose video or audio generation in this studio.
- Fireworks has no image-model catalog endpoint, so the model dropdown uses a curated list plus a free-entry field.
- Audio support was deprecated by Fireworks in mid-2026 and is not offered here.
Related docs
- Media Studios overview — shared render-queue and hand-off flow.
- Media Library — where finished images are saved.
- Settings — configuring AI providers.
Verified against v1.0.0