Skatter Me

The Process:

1. Take your photo
2. AI describes what it sees (Vision)
3. AI generates an image from that description
4. Repeat: describe → generate → describe → generate...
5. Watch your image evolve through 5 iterations!

Models Used:

LLaVA 1.5 7B Vision (image → text) SDXL Lightning Generation (text → image)

Vision Prompt:

Describe this image in vivid detail. Focus on the main subject, colors, style, mood, and composition. Be specific and artistic.

API Endpoints:

POST /api/gpu/vision/describe POST /api/gpu/image/generate