Skip to main content
LiveKit and Pipecat monitoring send conversation evidence from your agent through the Egma SDK. Retell monitoring imports completed calls with the Retell API. You do not need a test suite or a simulation run for any of these setups.

​
Start setup

  1. Open Agents and select Connect an agent.
  2. Choose Monitor production, then select Continue.
  3. Select LiveKit, Pipecat, or Retell and follow the setup prompts. For Pipecat, copy the prompt into a coding agent in your bot’s repository; it does the setup and asks you before it changes production.
  4. For LiveKit or Pipecat, add the SDK line and project key to your agent as described below. For Retell, connect your account and choose the voice agent to monitor.
  5. Complete a real conversation and check Traces for its evidence.
A LiveKit worker or Pipecat bot still needs the SDK setup; opening the wizard alone does not enable monitoring.

​
LiveKit

Follow Monitor a LiveKit agent to add the project key and SDK hook, verify a conversation, and stop monitoring when needed.

​
Pipecat

Follow Monitor a Pipecat agent to add the project key and the monitor line, verify a conversation, and stop monitoring when needed.

​
Retell

Follow Monitor a Retell agent to connect your account, enable call imports, and verify incoming conversations.

​
Review the evidence

Review production conversations to inspect transcripts, metrics, and grades.