whisper

Audio transcription via a REST API involves uploading common formats like OGG, MP3, or WAV to receive a job ID, polling the endpoint with health checks until processing completes, and then fetching the returned plain transcript text.

Speech-to-text conversion supports common ffmpeg-compatible audio formats including OGG/Opus, MP3, M4A, and WAV, making it suitable for transcribing WhatsApp voice notes and recorded meetings.

Asynchronous audio transcription handles long jobs by providing a job ID upon upload, allowing clients to poll with health checks, and enforcing a 60-second polling timeout while tracking status states like queued, processing, completed, or failed.

Yes, you can transcribe WhatsApp voice notes and meeting recordings because the service accepts common formats like OGG/Opus and M4A, processing them asynchronously to return plain English transcript text.

English audio transcription uses the faster-whisper medium.en model, which processes the audio asynchronously and can return either plain transcript text or full metadata depending on the API request.

The asynchronous transcription process limits polling to a 60-second timeout and currently focuses on English-language audio only, returning a failed status if processing exceeds the timeout or encounters incompatible input.