Whisper WebUI

A local speech-to-text browser app that creates subtitles with Whisper, translates subtitle text and supports speaker labeling under Apache 2.0.

Alternative to:

Whisper WebUI turns recordings into transcripts and subtitles through a browser interface running on your own machine. It's for people captioning videos or translating speech who want local audio processing.

Transcripts and subtitles: Process uploaded files, YouTube content or microphone recordings, then export SRT, WebVTT or plain text without timestamps.
Speech and text translation: Translate speech into English with Whisper, or translate subtitle files with Facebook NLLB models locally; the optional DeepL API sends subtitle text to an external service.
Audio preparation: Use Silero VAD to detect speech and UVR to separate background music before transcription.
Speaker labeling: Identify speaker changes with pyannote, whose model downloads require a Hugging Face token and acceptance of the model terms.
Whisper backend choice: Use the default faster-whisper backend for lower GPU memory use and faster transcription, or choose openai/whisper or insanely-fast-whisper, including compatible fine-tuned models from Hugging Face.

The app is open source under Apache 2.0 and supports Docker and Pinokio. A REST API backend lets other applications use it without the browser interface. YouTube input and Hugging Face model downloads use online services; the optional Google Colab notebook runs on hosted hardware.