Snowcut runs entirely in your browser. Your video is never uploaded. This page lists the other people's software the site serves to you and where to get its source code.
For video your browser cannot read itself (mostly iPhone HEVC in Chrome, Firefox and Edge), Snowcut downloads a converter
only when it is needed. It is the unmodified @ffmpeg/core and @ffmpeg/core-mt packages, version
0.12.10, from the ffmpeg.wasm project, which are FFmpeg built with
libx264 and libx265 and so are licensed under the GNU General Public License, version 2 or (at your option) any later
version. The wasm file is served cut into 16 MiB parts so the host accepts it; the bytes are unchanged. Snowcut's own
code talks to the converter only through command-line arguments and messages to a separate worker.
Licence texts: GNU GPL version 2, GNU LGPL version 2.1. Each archive below also carries its own licence files.
You may have the source of everything compiled into the converter, from this site, for as long as the site serves the
converter. The first archive holds the build recipe (Dockerfile and build/); the rest are the
libraries it lists, each at the exact commit named. For x264 and lame the upstream recipe names a
branch, not a fixed commit, so the commit shown is the one that branch pointed at on 2026-10-23.
| Component | Version | Commit | Download |
|---|---|---|---|
| ffmpeg.wasm | v0.12.10 | c3a763857c5e | ffmpeg.wasm-c3a76385.tar.gz, 5.1 MB |
| FFmpeg | n5.1.4 | 4729204c17f7 | FFmpeg-4729204c.tar.gz, 15.0 MB |
| x264 | branch 4-cores | 33cac6b77d5b | x264-33cac6b7.tar.gz, 0.9 MB |
| x265 | 3.4 | 2bb5520e9596 | x265-2bb5520e.tar.gz, 1.5 MB |
| libvpx | v1.13.1 | 10b9492dcf05 | libvpx-10b9492d.tar.gz, 5.6 MB |
| lame | branch master | 2badea1974ae | lame-2badea19.tar.gz, 1.5 MB |
| ogg | v1.3.4 | bada45718453 | ogg-bada4571.tar.gz, 0.3 MB |
| theora | v1.1.1 | 7ffd8b2ecfc2 | theora-7ffd8b2e.tar.gz, 1.1 MB |
| opus | v1.3.1 | e85ed7726db5 | opus-e85ed772.tar.gz, 0.8 MB |
| vorbis | v1.3.3 | 779816404319 | vorbis-77981640.tar.gz, 1.3 MB |
| zlib | v1.2.11 | cacf7f1d4e3d | zlib-cacf7f1d.tar.gz, 0.7 MB |
| libwebp | v1.3.2 | ca332209cb55 | libwebp-ca332209.tar.gz, 3.7 MB |
| freetype2 | VER-2-10-4 | 6a2b3e4007e7 | freetype2-6a2b3e40.tar.gz, 2.1 MB |
| fribidi | v1.0.9 | f9e8e71a6fbf | fribidi-f9e8e71a.tar.gz, 2.1 MB |
| harfbuzz | 5.2.0 | 4a1d891c6317 | part 1 + part 2, 33.5 MB (join the parts in order with cat) |
| libass | 0.15.0 | d149636f502f | libass-d149636f.tar.gz, 0.2 MB |
| zimg | release-3.0.5 | e5b0de6bebbc | zimg-e5b0de6b.tar.gz, 1.2 MB |
SHA-256 of every archive is in manifest.json.
| Software | Licence | Source |
|---|---|---|
| Mediabunny (reads, trims and writes video) | Mozilla Public License 2.0, unmodified | github.com/Vanilagy/mediabunny |
| @ffmpeg/ffmpeg (the JavaScript wrapper around the converter) | MIT | github.com/ffmpegwasm/ffmpeg.wasm |
| ONNX Runtime Web (runs the sound model) | MIT | github.com/microsoft/onnxruntime |
| YAMNet sound-recognition model, Google, as an ONNX export by Audio Magic | Apache License 2.0 (text). Changes: converted to the ONNX format; the weights are untouched. | huggingface.co/audiomagic/yamnet-onnx |
This is a record of what the site serves, not legal advice. Last checked 2026-10-23.