Clean up AI-generated pixel art into genuine pixel-perfect images.
Export pixel art as crisp SVG vector paths for scalable graphics.
Edit converted pixel art with built-in pencil, fill, and eraser tools.
Reduce pixel art color palettes to 64 or 32 colors.
| fulviuus/pixelloid | abhi-arya1/riff | andersondanieln/hexllama | |
|---|---|---|---|
| Stars | 13 | 13 | 13 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | hard | easy |
| Complexity | 2/5 | 4/5 | 2/5 |
| Audience | designer | developer | vibe coder |
Figures from each repo's GitHub metadata at analysis time.
Built with Tauri 2, so it requires installing Rust and Node.js toolchains to compile and run the desktop application locally.
Pixelloid is a desktop application that converts "pseudo-pixel art" into genuine pixel-perfect images. AI-generated pixel art often only looks like it has crisp pixels. In reality, those apparent pixels are made from many smaller source pixels softened by blurring, color noise, or inconsistent edges. Pixelloid detects the hidden grid in these images, creates an image at the grid's native resolution, and resolves every logical cell into one real output pixel. The entire process runs locally on your computer, so no images are uploaded to the cloud. The application offers several sampling modes to determine the final color of each pixel. Nearest mode matches a specific FFmpeg point sampling method. Medoid mode chooses a complete color sample observed inside each inferred cell. Smart mode compares nearest, medoid, and dominant source colors inside every logical cell, keeping the result that reconstructs the cell with the least destruction. All modes are deterministic and never generate new artwork. Optional palette reduction can reduce the final image to 64 or 32 colors after downscaling. Beyond conversion, Pixelloid includes a pixel editor with pencil, fill, eyedropper, eraser, crop, and zoom tools. Users can compare the source and result side by side, and overlay the original image to check edits. The app supports exporting the final result as an indexed PNG or an SVG vector file. The SVG export works from the final true-resolution result, merging pixels of the same color into crisp vector paths without embedding any raster image data. It can import PNG, JPEG, WebP, GIF, AVIF, and BMP files. Pixelloid is built with TypeScript, React, and the Tauri 2 framework. The project includes background removal, grid detection, and pixelization logic in its source library. It vendors the open-source unfake.js WebAssembly core as a secondary scale detector. The project is currently in alpha, with limits on source image size and background removal capacity. It is available under the Mozilla Public License 2.0.
A desktop app that turns fake pixel art into true pixel-perfect images. It detects the hidden grid, resolves colors, and exports clean PNG or SVG files, all running locally.
Mainly TypeScript. The stack also includes TypeScript, React, Tauri 2.
You can use and modify the code, but if you distribute changes you must share them under the same license.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly designer.
This repo across BitVibe Labs
Verify against the repo before relying on details.