ClipVuno uses these open-source components:

ffmpeg.wasm JavaScript wrapper 0.12.15: MIT license.
FFmpeg WebAssembly core 0.12.10: GPL version 2 or later (its compiled codec libraries include x264). Corresponding build scripts and source references: https://github.com/ffmpegwasm/ffmpeg.wasm/tree/main/packages/core and https://github.com/ffmpegwasm/ffmpeg.wasm/tree/main/build. FFmpeg source: https://ffmpeg.org/download.html. Core package: https://www.npmjs.com/package/@ffmpeg/core/v/0.12.10. Bundled executable engine is split into two byte chunks for hosting and reassembled unmodified at runtime.
heic2any 0.0.4: MIT license (bundled decoder includes libheif dependencies; see upstream repository notices). https://github.com/alexcorvi/heic2any

Optional local AI downloads (only after explicit user activation):
Transformers.js 3.8.1: Apache-2.0. https://github.com/huggingface/transformers.js
ONNX Runtime Web: MIT.
Xenova/whisper-tiny and Xenova/whisper-tiny.en: Whisper models, MIT.
Xenova/modnet: MODNet portrait-matting model, Apache-2.0.
Xenova/speecht5_tts and default HifiGAN vocoder: see model cards and upstream Microsoft SpeechT5 license at https://github.com/microsoft/SpeechT5.
Models and runtime assets are downloaded from Hugging Face and jsDelivr. User files are processed locally, not sent to these services.