Pydub

Convert WebM/Opus streams into 16kHz mono WAV files using Pydub and FFmpeg to meet OpenAI Whisper's precise input specifications.

This pipeline bridges the gap between browser-recorded WebM blobs and Whisper's 16kHz mono WAV requirement. Pydub provides the Pythonic interface while FFmpeg handles the heavy lifting: decoding Opus streams and resampling audio to the target 16,000Hz frequency. By standardizing your input to single-channel (mono) format, you ensure the ASR engine processes data without additional latency or conversion errors. It is a reliable, lightweight solution for developers building transcription features from MediaRecorder API outputs.

https://github.com/jiaaro/pydub