Run this workflow on InstaSD
Get started in minutes! Run this ComfyUI workflow online - no setup required.
Description
Turn your hand-drawn concepts into stunning final renders with the Sketch 2 Image SDXL workflow, tailored for clean and creative image transformations. This flexible setup supports any ControlNet preprocessor (like Canny, LineArt, Scribble, Depth, etc.) and lets you draw directly inside ComfyUI using the Canvas tab. Perfect for sketching characters, buildings, or abstract shapes and evolving them into highly detailed artworks using SDXL with ControlNet.
🎯 Features
- Draw Directly in ComfyUI – Use the Canvas tab to sketch your ideas without external tools.
- Supports Any ControlNet Preprocessor – Flexibly choose from Canny, Depth, LineArt, Scribble, and more.
- Interactive Resolution Controls – Preview and set resolutions before rendering.
- Multiple Comparison Stages – Includes preview and before/after sliders for results inspection.
- Dynamic Strength Settings – Fine-tune how closely the final image follows your sketch.
- Prompts Built-In – Comes preloaded with positive and negative prompts for high-quality output.
💡 Use Cases
- Character Design – Sketch and evolve unique character concepts into full illustrations.
- Architecture & Product Visualization – Draw structures or objects and bring them to life with ControlNet.
- Comic & Manga Creation – Turn inked panels or line art into fully rendered scenes.
- Concept Art – Rapid prototyping of visual ideas directly in ComfyUI.
- AI-Assisted Sketch Refinement – Perfect for turning rough outlines into polished digital art.
⚙️ How It Works
- Open the Canvas Tab – Start with a blank canvas or load an image and draw over it.
- Select Aspect Ratio & Resolution – Use the resolution nodes to fit your canvas needs.
- Pick a Preprocessor – Choose any ControlNet-compatible preprocessor like Canny or Scribble.
- Load a ControlNet Model – Example:
diffusion_pytorch_model_promax.safetensors. - Enter Prompts – Add your positive (what you want) and negative (what to avoid) prompts.
- Adjust Control Strengths – Control creativity vs. fidelity with strength and end_percent sliders.
- Run KSampler – Generates your image based on all settings.
- Preview & Compare – Use comparison tools to inspect results before saving.
- Save Final Image – Export your masterpiece or copy it directly into another ComfyUI flow.
Credits: pixaroma
Models
Nodes
ImageScalePreviewImageSaveImageeasy showAnythingLabel (rgthree)Image Filter AdjustmentsFluxResolutionNodeImage Comparer (rgthree)NoteImage ResizeCheckpointLoaderSimpleCLIPTextEncodeEmptySD3LatentImageControlNetLoaderControlNetApplyAdvancedKSamplerVAEDecodeAIO_PreprocessorVAEEncodeFast Groups Muter (rgthree)LoadImageCanvas_Tab