What is sprite-lab?

Sprite Lab is a browser tool for removing image backgrounds and slicing spritesheets into individual frames. Everything runs locally in the browser, nothing is uploaded to a server, and it does not use any AI models. The tool detects what kind of background an image has, such as magenta, green screen, checkerboard, or a solid color, and removes it automatically. It uses a multi pass alpha matting process that estimates how much of each edge pixel belongs to the foreground versus the background, then cleans up the color so leftover halos and fringes around the sprite are eliminated instead of just cut and pasted. Once the background is gone, the app can slice a spritesheet into separate frames. Rather than cutting the image into a fixed grid, it scans for groups of connected pixels to find where each sprite actually is, so sheets with irregular frame sizes or spacing still get split correctly. Users get an animation preview with adjustable playback speed, can click individual frames to exclude them, and export everything as a ZIP file containing each frame, a single combined strip image, and a manifest file that lists frame dimensions and positions. To run it, you clone the repository, install dependencies with npm, and start a local development server with npm run dev, then open the URL it prints in a browser. For deployment, running the build command produces a static output folder that can be hosted anywhere, including GitHub Pages, Netlify, or Vercel, since there is no backend server involved. Under the hood, the project is built with Vite and TypeScript, uses the browser's Canvas API to do the actual pixel work, and relies on JSZip for packaging exports. It also has a set of command line tests that run the same processing pipeline outside the browser to check specific cases like checkerboard backgrounds and light colored halos. The project is released under the MIT license.