Scribble Diffusion

Scribble Diffusion generates images from rough sketches. Draw a quick doodle in the browser, and the AI analyzes your lines and turns them into a finished picture.

How it works

Scribble Diffusion is built on the ControlNet extension for Stable Diffusion. The model detects key points and contours in your sketch and uses them as guidance to produce digital art.

Using Scribble Diffusion

The tool runs online and doesn’t require registration or a paid subscription. To use it, you’ll need to provide a Replicate API token, which unlocks the drawing canvas.

• Draw directly on the virtual canvas in your browser
• Use undo and clear controls while sketching
• Add a text prompt to steer the style or subject for more accurate results

Output and key features

Generation typically takes about a minute. The finished image can be downloaded as a PNG.

• Sketch-to-art image generation
• Open-source project
• English-language interface
• Sketching is available only online
• PNG downloads