Turn the speech into word-by-word captions. The spoken word lights up in your accent colour.
Choose an audio file in section 1 first.
Runs on this device. The model downloads once from Hugging Face and is cached; your audio is never uploaded.
Claude Code and Codex can't hear audio, but they can run Whisper on your computer for you, which is more accurate for long or noisy recordings.
- Copy the prompt and paste it into Claude Code or Codex, running on the computer that has the audio file.
- It transcribes locally with Whisper and saves a
.captions.json file next to the audio.
- Import that file here.
SRT and WebVTT files from any other tool work too.