Licenses
This site is a front end for other people's work. Here is whose.
FFmpeg
All processing is done by FFmpeg, compiled to WebAssembly. The build shipped here is configured with --enable-gpl and includes libx264, libx265, libvpx, libmp3lame, libtheora, libvorbis, libopus, libwebp, libfreetype, libfribidi, libass, libzimg, so it is distributed under the GNU General Public License, version 2 or later.
The unmodified binary served by this site is at /assets/vendor/ffmpeg-core.wasm, with its loader at /assets/vendor/ffmpeg-core.js. Corresponding source is published by the FFmpeg project and by the ffmpeg.wasm project below. This site is not affiliated with or endorsed by the FFmpeg project.
ffmpeg.wasm
The WebAssembly build and its JavaScript wrapper come from ffmpeg.wasm (wrapper 0.12.15, core 0.12.10). The wrapper is MIT licensed; the core carries FFmpeg's license.
Whisper
Speech recognition uses Whisper, released by OpenAI under the MIT license. The site loads the ONNX conversions published by the onnx-community organisation. This site is not affiliated with or endorsed by OpenAI.
The image models
The explicit-content check uses vit-base-nsfw-detector by AdamCodd. The age estimate uses two: detr-face-detection to find the faces, then age-gender-prediction-ONNX, an ONNX conversion of work by abhilash88, to read each cropped face. All three are under the Apache License 2.0. None of their authors is affiliated with or endorses this site, and their accuracy is their claim, not a promise made here.
Transformers.js and ONNX Runtime
The model is loaded and executed by Transformers.js (Apache License 2.0) running on ONNX Runtime Web (MIT license). Both are served unmodified from /assets/vendor/.
DejaVu Sans
The font used by the text and subtitle tools is DejaVu Sans, under the DejaVu Fonts license, a permissive Bitstream Vera derivative.
This site
The page markup, styling and tool definitions are original work. The trademarks and names of the projects above belong to their respective owners.