Dev.to
7/1/2026

AetherCut Hardware acceleration.
Short summary
AetherCut processes video with GPU acceleration and runs Whisper transcription locally via WebAssembly, supporting 14 languages entirely on-device. Hardware acceleration delegates to GPU; models cache after first download. All processing stays local—zero cloud transmission.
- •GPU-accelerated video processing via 'prefer-hardware' flag for Chrome 94+ and Edge
- •Local Whisper transcription in 14 languages via WebAssembly; models cached after first download
- •Zero cloud transmission—all audio and video processing stays on-device
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



