Compix — Compare, Crop and Stage AI Images in Your Browser
Every AI tool generates new images. Compix combines existing ones. Upload your Midjourney, Stable Diffusion, or Flux outputs. Draw freeform shapes on the parts you want to swap: head, outfit, accessories. The combinatorial engine generates every possible combination. 9 images × 1 shape = 72 unique variants. 9 images × 3 shapes = 6,500+ combinations. No AI model runs. No GPU needed. No credits spent.
How It Works
Load your AI-generated character images. One becomes the anchor: your base. Draw freeform shapes around the regions you want to swap: the head, the outfit, the accessories. Inside each shape, instantly see that exact region from every other loaded image. The combinations compound: images × (images^shapes − 1). With 9 images and 3 shapes: 6,552. The engine computes all of them in seconds, at full resolution, pixel-perfect.
Why This Exists
AI generation platforms make money when you generate more images. When your image is 90% right and you regenerate, that's revenue for the platform. Compix extracts and recombines the best parts from images you already generated. Generation platforms will never build this because it would cannibalize their revenue model.
Also Included
Blink comparison for detecting character drift between generations. Pixel diff heatmap showing exactly which pixels changed. Split wipe for precise pixel-level inspection. Layer compositor with snap guides and animation timeline. Export as PNG, GIF, or MP4. All processing runs in your browser — nothing is uploaded to any server. Works offline.
Please enable JavaScript to use this application.
