Transcribe

Transcribe audio files to text using Whisper.

Process

1. Check audio format - Ensure file is in supported format (WAV, MP3, etc.)
2. Convert if needed - Use ffmpeg to convert to WAV if necessary
3. Choose model - Use smaller model for speed, larger for accuracy
4. Run transcription - Use Whisper CLI or Python interface
5. Extract text - Get the transcribed text output
6. Review quality - Check transcription accuracy and fix obvious errors

Examples

# Whisper CLI (preferred if installed)
whisper /path/to/audio.wav --model small --output_format txt
cat /path/to/audio.txt

Python fallback

python3 - << 'PY'
try:
    import whisper
except Exception:
    raise SystemExit("whisper not installed")

model = whisper.load_model("small")
result = model.transcribe("/path/to/audio.wav")
print(result.get("text", ""))
PY

Tips

• Use a smaller model for speed.
• Convert audio to WAV if needed: ffmpeg -i input.mp3 output.wav.