Sending Audio
1 min
The expected audio type for Autoscriber to process is:
- Base64 encoded audio/pcm
- Format: 16-bit PCM, little-endian
- Sample Rate: 16,000 Hz
- Channels: 1 (mono)
The recommended audio chunk interval is:
- Send a chunk every 10 seconds
For better transcription quality, Autoscriber doesn't transcribe every chunk individually - the chunks are concatenated and return results intermediately per X chunks (the value of X is context dependent).