tidy

no description on the landing page

private fitting backend: no public editor UI, 45s GPU request

Pro-Realism (Qwen-Image-Edit-2511 + AIO v18), default GPU slot

xlarge is required: the default slot aborts for anonymous callers

Test the default ZeroGPU slot at 60s

xlarge slot at 60s; xlarge counts double against the budget

Request an xlarge ZeroGPU slot, as the 57.7GB transformer needs

Give infer a 120s GPU allocation

Add a plain gr.api endpoint that returns errors in the response

Pin dependencies so the build resolves

Qwen-Image-Edit Rapid AIO with LoRAs

Capture the caller's dimensions before the canvas resize

Return result at the caller's dimensions; revert the API signature change

FireRed-Image-Edit with optional negative prompt and caller-sized output

Update requirements.txt

Update requirements.txt

Update requirements.txt

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update README.md

Update README.md

Update README.md

Update requirements.txt

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update utils_mask.py

Update utils_mask.py

Update utils_mask.py

Update utils_mask.py