LAB QUOTA · OK
[ img-forge:// ] experimental
cat: image model: @cf/bytedance/stable-diffusion-xl-lightning

Text-to-image via SDXL Lightning. 4-step generation finishes in ~3 seconds. Saves to your R2 gallery (sign in required). 5 images/day on free tier.

// system prompt
Image generation prompt. The UI will append safety/quality modifiers automatically:
  ", high detail, professional lighting, sharp focus, photorealistic"

Users can edit those via Advanced. Negative prompts default to:
  "lowres, blurry, watermark, text, error, jpeg artifacts, mutated hands"

Returns: image/png, 1024×1024.
⚡ Cloudflare Workers AI · quota deducted on success
// sample output
[STUB]

When Workers AI is bound, /lab/api/ai/image will return a PNG
rendered by SDXL Lightning in 4 steps (~3s wall time).

The tool's React island will show:
  - Live progress bar
  - 4-image grid (4 random seeds)
  - Save-to-gallery / download buttons
  - "Vary this one" button → posts back to /api/ai/image with seed
// powered by cloudflare workers ai · quota deducted on success ← back to catalog