ComfyUI - Coloring Book using Flux Kontext

Use this ComfyUI workflow to convert an input image into a coloring-book-style image, with preview and saved-image outputs, using the Flux.1 Kontext model.

The workflow includes nodes for loading and encoding a reference image, applying text conditioning and sampling, and previewing, comparing, and saving images.

Provide an image with LoadImage; the workflow uses ReferenceLatent, FluxKontextImageScale, and easy imageSize in the image-processing path.

View a result with PreviewImage, compare images with Image Comparer (rgthree), and save an image with SaveImage.

Open the Civitai listing at https://civitai.com/models/1851404.

Node packs listed are bjornulf_custom_nodes 1.1.8, comfyui_fill-nodes 1.6.3, comfyui-easy-use 1.3.1, and rgthree-comfy 1.0.2508012353.

Model files listed include ae.safetensors, clip_l.safetensors, and FLUX-Kontext/colossusProject_v10FP8UNETONLY.safetensors.

They also include FLUX/colossusProjectFlux_v10BehemothAIOFP16.safetensors, FLUX1/ae.safetensors, and t5xxl_fp8_e4m3fn_scaled.safetensors.

Style options include southpark animation, claymation, disney princess cartoon, japanese woodblock carving, pixel art, Vaporwave, and Zen brush illustration.

Other style options include Underground comix, Psychedelic trip, Stencil art, Encaustic pigmented wax, thick texture Impasto, Acrylic pour cell marbling, and dark evil creepy style.

Additional options include plushie style, everything is soft furry plush fabric; action figure toys, everyone is an action figure toy and items are action toy accessories; ivory carving; wood carving; sand sculpture; steampunk clockwork and gears; and neonpunk.

Before running it, verify that every listed model file is available under the exact name shown.

Suggestion · not verified

No license is listed for this workflow; check the Civitai terms before reuse.

Suggestion · not verified