Prompt-based AI inpainting
Select an area of an image and describe the desired change with a prompt, producing a generated edit for that region.
Scumble
Freemium
Free, open-source desktop image editor for AI-assisted inpainting with layers and PSD export.
O que é Scumble?
Scumble is a free, open-source desktop image editor focused on AI inpainting. Its central workflow is to select an area, enter a prompt, and generate a colour-matched layer for the edit. The project combines generative image editing with a layered document workflow and supports PSD export for continued work in compatible software.
The repository lists FLUX 3 Image, FLUX.2, GPT Image, Nano Banana, Seedream, and Qwen as model options. These can be used through an API key or through the user’s own ComfyUI setup. Scumble also advertises MCP support, although the available source does not describe the full command set or integration procedure.
O que Scumble pode fazer?
Colour-matched layers
The project is designed to place generated inpainting results as colour-matched layers, supporting further adjustment within the document.
Multiple model options
The repository names FLUX 3 Image, FLUX.2, GPT Image, Nano Banana, Seedream, and Qwen as supported model options.
API key or ComfyUI workflows
Connect to supported providers with an API key or use your own ComfyUI setup as the image-generation backend.
Layered and PSD-based editing
Work with layers in the editor and export documents in PSD format for continued editing in compatible image applications.
MCP support
Scumble includes MCP support, providing a route for use with compatible Model Context Protocol tooling; detailed capabilities are not specified in the available source.
Casos de uso
“Targeted image corrections”
Select an unwanted or incomplete area, describe the replacement, and use inpainting to make a localized change without rebuilding the whole image.
“Layered generative exploration”
Try alternate AI edits while keeping generated results in a layer-oriented document workflow, making it easier to compare or refine variations.
“ComfyUI-based image work”
Use an existing ComfyUI setup as the generation backend when a user prefers to work with their own ComfyUI environment rather than only an API-key connection.
“PSD handoff”
Export layered work as PSD when the image needs to move into another compatible editor for additional compositing or finishing.
Perguntas frequentes
What is Scumble?
Scumble is a free, open-source desktop editor for AI-assisted image inpainting. Its main workflow is selecting an area, entering a prompt, and generating a colour-matched layer.
Which image models does Scumble list?
The repository lists FLUX 3 Image, FLUX.2, GPT Image, Nano Banana, Seedream, and Qwen. The available source does not provide detailed capability or version comparisons for these options.
Can Scumble use my own ComfyUI setup?
Yes. The project description says it can use a user’s own ComfyUI, in addition to connections made through an API key.
Can I continue editing the result elsewhere?
Scumble advertises layers and PSD export, so layered work can be exported for continued editing in compatible PSD software. The source does not specify export limitations or compatibility details.
Informações rápidas
- Category
- AI image editing and desktop design tool
- License
- GPL-3.0
- Core workflow
- Select an image area, prompt an edit, and generate a colour-matched layer
- Model options named by the project
- FLUX 3 Image, FLUX.2, GPT Image, Nano Banana, Seedream, and Qwen
- Backend options
- API key or the user’s own ComfyUI setup
- Output and automation features
- Layers, PSD export, and MCP support
Alternativas ao Scumble
CapCut
capcut.com
A cross-platform video and image editor for creators, marketers, and teams.
Polarr
polarr.com
AI-assisted photo editing, reusable filters, and selective adjustments across mobile and desktop photo tools.
Bylo AI
bylo.ai
Gerador de imagens e editor de fotos com IA na web, com GPT-4o, GPT-Image-1 e Nano Banana
AI Image Generator
ai-img-gen.com
AI Image Generator é uma ferramenta web para criar imagens geradas por IA a partir de prompts e explorar exemplos de resultados. Também inclui um editor de fotos integrado para fazer ajustes básicos após a geração.