Supervisor controls let your team monitor and intervene in active AI calls without disrupting the caller experience. Three modes are available: listen, whisper, and barge.
Each mode returns a connection token you use to join the live call audio session.
Modes at a Glance
Mode
You can hear
You can speak to
Caller hears you
You appear in call
Listen
Agent + caller
Nobody
No
No (hidden)
Whisper
Agent + caller
Agent only
No
No (hidden)
Barge
Agent + caller
Everyone
Yes
Yes
Listen
Join a call as a hidden observer. Use this for quality assurance, training review, or compliance monitoring.
Code
curl -X POST https://api.neuratel.ai/v1/voice-sessions/{call_id}/listen \ -H "Authorization: Bearer nk_live_your_api_key"
Code
import httpxresp = httpx.post( f"https://api.neuratel.ai/v1/voice-sessions/{call_id}/listen", headers={"Authorization": "Bearer nk_live_your_api_key"}).json()print(resp["mode"]) # "listen"print(resp["token"]) # connection tokenprint(resp["server_url"]) # media server URL
The token in the response authenticates your connection to the live call audio session. Use it with the Neuratel web SDK or compatible WebRTC client to join the audio.