Video to GIF Converter

A GIF is a genuinely old, tightly specified image format: it can only address up to 256 colors at a time, and every frame is compressed with LZW, the same lossless algorithm classic UNIX "compress" used. Converting a full-color video clip into that format means solving two real problems — picking the best possible 256-color (or fewer) palette for the clip, and compressing each frame's pixel data losslessly against that palette — both handled here with algorithms implemented directly against the GIF89a specification.

Color reduction uses median-cut quantization: pixels sampled across the whole clip are recursively split into buckets along whichever color channel has the widest range, until the palette reaches the size you chose, with each bucket's average color becoming a palette entry. Optional Floyd–Steinberg dithering then scatters the unavoidable rounding error from each pixel into its neighbors, which is what keeps skies, skin tones, and other smooth gradients from banding visibly with a limited palette.

Compression uses GIF's own LZW variant — variable-width codes starting at the palette's minimum bit depth, a dictionary that grows as repeated pixel sequences are found, and the specific LSB-first bit-packing GIF requires. This isn't an off-the-shelf compression library; it's a direct implementation of the algorithm described in the GIF89a spec, verified by decoding its own output back with a real, independent GIF decoder and confirming every pixel matches — including edge cases like the dictionary filling up mid-frame and needing to reset.

Frame extraction happens by seeking your video element to evenly spaced timestamps across the trim range and drawing each one to a canvas at your chosen output size, entirely through browser APIs. Nothing about your video — the file itself, the frames extracted from it, or the finished GIF — is ever sent anywhere; the whole pipeline from upload to download runs in this browser tab.