Audio encoding - AssemblyAI

The encoding determines the sample rate and bit depth of the audio going in (microphone) and coming out (agent speech). Set it on the agent under input.format and output.format when you create or update it, or inline via session.update. Most agents can leave this alone — the defaults are the highest quality. Change it mainly for telephony. For how to actually stream and play the audio bytes, see Stream audio. Input and output encodings are independent and can differ. Both default to audio/pcm (24 kHz) if omitted.

Encoding	Sample rate	Best for
`audio/pcm`	24,000 Hz	Default. Highest quality, ideal for browser and app use.
`audio/pcmu`	8,000 Hz	Telephony (G.711 μ-law).
`audio/pcma`	8,000 Hz	Telephony (G.711 A-law).

For telephony, use audio/pcmu or audio/pcma (8 kHz) to match the phone network and avoid resampling. See Connect to Twilio for a full phone integration. Set format.encoding under input and output. You can also pass an explicit sample_rate inside format:

curl -X POST https://agents.assemblyai.com/v1/agents \
  -H "Authorization: $ASSEMBLYAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Support Assistant",
    "system_prompt": "You are a friendly support agent. Keep replies under two sentences.",
    "voice": { "voice_id": "ivy" },
    "input":  { "format": { "encoding": "audio/pcmu", "sample_rate": 8000 } },
    "output": { "format": { "encoding": "audio/pcmu", "sample_rate": 8000 } }
  }'

Field	Type	Required	Notes
`input.format.encoding`	string	No	`audio/pcm`, `audio/pcmu`, or `audio/pcma`. Default `audio/pcm`.
`output.format.encoding`	string	No	Same values as input. Default `audio/pcm`.
`format.sample_rate`	integer	No	Sample rate in Hz. Determined by the encoding if omitted.

Set the volume separately — see Volume. When configured inline via session.update, output.format is immutable after session.ready and must be set on your first update.