Added model upscaling for much better results
Runtime profile
Remaining hardware or download-size estimates are not available yet.
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Use Simple Kontext Inpaint Cropping workflow v2.0 with Flux.1 Kontext to prompt edits to a selected part of an image: provide an image and mask the area to change, and create an edited image.
The workflow includes InpaintCropImproved and InpaintStitchImproved nodes, plus PreviewImage and SaveImage nodes.
For focused edits, the creator says it works best with a LoRA that changes only one small part of the image.
Suggestion · not verified
To select an area, right-click the Load Image node, choose Open Mask Editor, and mask the part you want to change.
The included versions list these node packs: comfyui-inpaint-cropandstitch, comfyui-kjnodes, and rgthree-comfy; the Nunchaku version also lists ComfyUI-nunchaku.
Across the two versions, listed model files are ae.safetensors, clip_l.safetensors, Marksketdra-4bits.safetensors, RealESRGAN_x4plus.safetensors, svdq-int4_r32-flux.1-kontext-dev.safetensors, and t5xxl_fp16.safetensors for the Nunchaku version; the non-Nunchaku version lists flux
Version 2 adds model upscaling for the masked area and includes Nunchaku and non-Nunchaku versions.
The creator suggests adjusting the inpaint crop node and placing an upscaler after cropping but before resizing for Flux Kontext; final upscalers mainly make the image larger.
Suggestion · not verified
The creator notes that Flux Kontext often changes the resolution slightly, which can leave the final image misaligned with the original; check alignment after editing.
Suggestion · not verified
If you were among the first 10 people to download v2, the creator advises downloading it again or replacing the Load Image (from outputs) node with a regular Load Image node.
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
3 excerptsSource context: 591 downloads · Type Workflows · Base model Flux.1 Kontext
Requirements
VAE · Hugging Face · Comfy-Org/Lumina_Image_2.0_Repackaged · Lumina Image 2.0 Repackaged
Text encoder · Hugging Face · comfyanonymous/flux_text_encoders · Flux Text Encoders
Marksketdra-4bits.safetensors
Not resolvedLoRA · Unknown
Upscaler · SAFETENSORS · Unknown
svdq-int4_r32-flux.1-kontext-dev.safetensors
Not resolvedCheckpoint · Unknown
Text encoder · 9.12 GB · SAFETENSORS · Unknown
comfyui-inpaint-cropandstitch
FoundNode pack · Registry
Node pack · Registry
Source context: 544 downloads · Type Workflows · Base model Flux.1 Kontext
Source context: 531 downloads · Type Workflows · Base model Flux.1 Kontext
Node pack · Registry
rgthree-comfy
FoundNode pack · Registry