Acknowledgements
BaoCut is possible because people shared the speech, machine-learning, media, and design tools their own work was built on.
Core projects
| Project | Contribution |
|---|---|
| speech-swift | Foundation for BaoCut’s local speech and speaker pipeline. |
| MLX Swift | GPU-accelerated machine learning on Apple silicon. |
| Swift Transformers | Model Hub downloads and transformer tooling. |
| FFmpeg | Media inspection, conversion, and export. |
| yt-dlp | Importing media from supported web links. |
Models and design
BaoCut’s local pipeline builds on Qwen3-ASR, Whisper, Silero VAD, pyannote, and WeSpeaker. The interface follows Adobe Spectrum 2 and uses open-source typefaces from Adobe Fonts and independent type designers.
Each project and model remains the work of its authors and is provided under its own license and terms.