ComfyUI user says SDXL inpainting keeps leaking outside the mask and ignoring prompts
WallabyFearless7863 · reddit · 2026-07-24
The author describes trouble getting inpainting to work reliably in ComfyUI with SDXL/Anima: the mask often stays visible, outputs ignore the prompt, and the generated area spills outside the mask. They share example results, screenshots of the workflow, and ask whether the issue is model-specific or caused by settings, masking, or workflow design.
The post is essentially a troubleshooting request for image inpainting, with examples showing poor edge cleanup and inconsistent edits. It highlights mask feathering, blur, and node separation attempts, but the results still fail to match a reference tutorial.
More from Multimodal
- Android NPU demo generates Stable Diffusion 2.1 images locally in 6.1 seconds — nntb · 2026-07-24
- Stop making cinematic clips! Dev calls for creative FLUX3 video prompts — cocktailpeanut · 2026-07-24
- UnrealTwin animates static 3D Gaussian splats in the browser with GPU shaders — danbri · 2026-07-24
- AMD 7900 XTX user trains FLUX.2 LoRA on 24GB VRAM with native ROCm — elderon_echar · 2026-07-24
- FLUX Video arrives with a demo and workflow notes from Gossip Goblin — DavidmComfort · 2026-07-24
- Early FLUX 3 Video test shows a mech fight, a kaiju, and the Seattle Space Needle — DavidmComfort · 2026-07-24