WhisperX Gradio Interface

Transcribe audio files with word-level timestamps using WhisperX.

WhisperX adds word-level timestamps to Whisper's transcriptions through phoneme-level alignment with a fine-tuned speech recognition model. It also supports speaker diarization.

Notes

  • The transcription process may take a while depending on the model size and audio length
  • For diarization to work, you may need to set the HF_TOKEN environment variable
  • For custom alignment models, provide the model name or path
Whisper Model
Task
Device
Compute Type