Remove temporary diff file

Fix Demucs dependency issue by adding torchcodec and improve subprocess error handling

Refactor backend: Replace standard Whisper with faster-whisper, add Demucs vocal separation and WebRTCVAD for improved alignment

Remove temporary diff output file

Add /translate endpoint for SRT translation via OpenAI

Update server.py and requirements.txt

Fix 500: Add WhisperX compat wrappers

Fix VAD startup crash: Enforce silero VAD and register OmegaConf safe globals

Fix import: Replace unavailable whisperx.utils.write_srt with internal implementation

Update dependencies and server configuration

Upload 3 files

initial commit