# AI use cases for Scumble (DenRakEiw/scumble) Repo: [DenRakEiw/scumble](https://github.com/DenRakEiw/scumble) | Use cases: 4 | Page: https://dedupedai.com/use-cases/denrakeiw-scumble/ Scumble is a free, open-source Windows desktop image editor that uses AI models to inpaint, edit, upscale and generate, with each result returned as its own layer. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Replace part of a photo with an AI layer that matches: https://dedupedai.com/use-cases/denrakeiw-scumble/replace-part-of-photo-with-colour-matched-layer/ - Generate a new picture from boxes and reference images: https://dedupedai.com/use-cases/denrakeiw-scumble/place-boxes-and-references-into-new-image/ - Let Claude Code edit pictures in Scumble through MCP: https://dedupedai.com/use-cases/denrakeiw-scumble/let-claude-code-edit-pictures-through-mcp/ - Export a layered PSD from a local ComfyUI edit: https://dedupedai.com/use-cases/denrakeiw-scumble/export-layered-psd-from-local-comfyui/ ## How to replace part of a photo with an AI layer that matches with Scumble Repo: [DenRakEiw/scumble](https://github.com/DenRakEiw/scumble) | Difficulty: Beginner | Time: 30 minutes | Basis: From the README Select an area, describe the change, and get a colour-matched layer you can mask, blend or undo. **Who it is for:** A photographer or social media editor who needs to remove an object or add something to one picture and has only an API key, no ComfyUI. **The problem:** Retouching a region with AI usually returns a flat new image that does not match the lighting, and changing it later means redoing the whole edit. **The outcome:** The edit sits as its own layer over the selection, faded at the edge and colour-matched to the surrounding picture, and can be adjusted or flattened when finished. ### Steps 1. Download the Windows installer from https://github.com/DenRakEiw/scumble/releases/latest or run `winget install scumble` in a terminal. 2. Open Scumble, press Ctrl+, to open Settings, go to API providers, paste your provider key and press Save. 3. Open your picture, select the area to change with the brush, rectangle, lasso or magic wand, and type what should be there. 4. Run the generation and check the new layer; move the Match slider in the layer's row if the colour does not blend well. 5. Mask or erase parts of the layer as needed, then flatten the document only when you are happy with the result. ### Prompt to paste into your agent ```text Using the Scumble repo at https://github.com/DenRakEiw/scumble, help me replace the selected area of with through my API key for . Expected output: an edited picture with the result as a separate layer. Verify that the result appears as its own layer and that the status line names the provider used. ``` ### Caveat Scumble is at version 0.1.x, and the README says the API routes for many providers have not yet been run against live services, so test with one provider first. Risk, credentials: Your provider API key is billed directly by that provider, so keep it private and check the provider's usage limits before running many edits. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/denrakeiw-scumble/replace-part-of-photo-with-colour-matched-layer/ ## How to generate a new picture from boxes and reference images with Scumble Repo: [DenRakEiw/scumble](https://github.com/DenRakEiw/scumble) | Difficulty: Beginner | Time: 30 minutes | Basis: From the README Draw boxes on an empty canvas, describe each one, and have FLUX 3 Image build the scene. **Who it is for:** A content creator planning a composed image who wants to say where each element goes instead of writing one long prompt. **The problem:** Text prompts alone give little control over where people, objects and words end up in a generated scene. **The outcome:** A generated picture where each boxed area follows its description, with reference layers named @img1 and @img2 in the prompt and returned as layers. ### Steps 1. Install Scumble from the Microsoft Store with `winget install scumble`, or use the installer from the latest GitHub release. 2. Open Settings with Ctrl+, , go to API providers, and paste your Black Forest Labs key for FLUX 3 Image, then press Save. 3. On an empty canvas, draw a box for each element and describe it as something new, something kept, moved or removed, or words to render. 4. Add a reference picture as a layer and name it in the prompt as @img1, then choose Generate new. 5. Check the result layer, adjust the Match slider if needed, and export when the scene is correct. ### Prompt to paste into your agent ```text Using the Scumble repo at https://github.com/DenRakEiw/scumble, create a new image on an empty canvas with boxes described as using FLUX 3 Image with my Black Forest Labs key. Expected output: a generated picture returned as a layer. Verify that each boxed area reflects its description. ``` ### Caveat The README says boxes for FLUX 3 Image work on every provider Scumble runs it on, but the Black Forest Labs route is the one it confirms ran live with boxes. Risk, credentials: The Black Forest Labs key is billed directly to your account, so store it only in Scumble's settings where the system credential store keeps it. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/denrakeiw-scumble/place-boxes-and-references-into-new-image/ ## How to let Claude Code edit pictures in Scumble through MCP with Scumble Repo: [DenRakEiw/scumble](https://github.com/DenRakEiw/scumble) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Connect Claude Code to Scumble's MCP server so an agent can select, generate, match and export. **Who it is for:** A developer or designer who already works in Claude Code and wants an agent to run repetitive image edits in the same window they can watch. **The problem:** Batch image edits are slow to do by hand, and an agent without access to the editor cannot see the picture or use its undo history. **The outcome:** Claude Code opens, selects, generates, matches and exports pictures through Scumble's MCP tools, with the same undo as manual work. ### Steps 1. Install Scumble and open it, then choose Help, then Copy MCP registration. 2. Paste the copied registration into your Claude Code MCP configuration. 3. Open the picture you want to work on in Scumble so the agent has a document to act on. 4. Ask Claude Code to select the area and run a generation, then check the layer that comes back. 5. Export the document from Scumble or ask the agent to export it, and review the result in the window. ### Prompt to paste into your agent ```text Using the Scumble repo at https://github.com/DenRakEiw/scumble, connect through its MCP server and select the area in the open document, then generate . Expected output: a new layer in the open document and an exported file at . Verify that the new layer appears in the window and that undo reverts it. ``` ### Caveat The README describes 114 MCP tools, but its own status notes say the MCP server has been run live while other agent routes have not, so check each tool's result before relying on it. Risk, credentials: An agent that runs generation through your API keys will spend provider credit, so set limits on the agent's task and keep the keys out of prompts. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/denrakeiw-scumble/let-claude-code-edit-pictures-through-mcp/ ## How to export a layered PSD from a local ComfyUI edit with Scumble Repo: [DenRakEiw/scumble](https://github.com/DenRakEiw/scumble) | Difficulty: Advanced | Time: 2 hours | Basis: From the README Run inpainting on your own ComfyUI and export the result as a PSD with layers and masks. **Who it is for:** A retoucher or studio artist with a local ComfyUI setup who needs layered files for a client's graphic design software. **The problem:** Local ComfyUI outputs are flat images, so a client who wants to adjust edits in Photoshop needs the layers kept. **The outcome:** A PSD file with each AI edit as a separate layer and its masks kept, produced from a recipe that runs on your own machine. ### Steps 1. Open a terminal, go to your ComfyUI custom_nodes folder, and run `cd ComfyUI/custom_nodes`. 2. Run `git clone https://github.com/DenRakEiw/ComfyUI-InpaintCanvas`, then restart ComfyUI. 3. In Scumble's top bar, type your ComfyUI server address and press Connect, then choose a recipe that uses the Inpaint Canvas node. 4. Make your selections and edits, then export to PSD with layers and masks from Scumble's export options. 5. Open the PSD in your design software and confirm each edit is a separate layer with its mask. ### Prompt to paste into your agent ```text Using the Scumble repo at https://github.com/DenRakEiw/scumble and its ComfyUI-InpaintCanvas node pack, run an inpainting edit on against my local ComfyUI at and export a PSD to . Expected output: a PSD file with one layer per edit. Verify that the PSD opens with layers and masks intact. ``` ### Caveat The README says the ComfyUI-InpaintCanvas pack is not in the ComfyUI Manager yet and that Scumble will not install it for you, so the git clone step must be done by hand. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/denrakeiw-scumble/export-layered-psd-from-local-comfyui/