Access Voice AI agent conversations, recordings, transcripts, and execution data. Monitor performance metrics, debug with logs, and export data for analysis.
Call History (Agent Conversations) displays all historical conversations with your agents. View performance metrics, listen to recordings, read transcripts, and access raw execution data for debugging and analysis.
Click Recordings, transcripts, etc to view the full conversation data.
Conversation data panel
Section
Description
Recording
Audio waveform with play, copy, and download options
Transcript
Full conversation showing Assistant and User messages
Summary & extracted data
Call summary and extraction results, when the agent has them configured
The Recording card only appears for calls that produced audio. Text chats from Chat with agent are logged as websocket outbound executions with a transcript but no recording.
Use the copy button to quickly copy the recording URL or transcript text.
Click the Debug (bug) icon in a call row to open the debugging view for that execution in a new tab. You can also open Monitor → Debugging in the left navigation menu and paste an execution ID into the search box to look up any call.
Debugging view
The top of the page shows a recording player, the call type, status, webhook delivery status, batch, initiated time, hangup by, and hangup reason. Use Download recording to save the audio and Rate call to submit feedback. Click Hide details to collapse the call details.
Tab
Description
Summary
Agent and telephony configuration used for the call, function tools, and cost breakdown
Timeline
Latency per Turn chart (transcriber finalize, transcriber → LLM, LLM, time to first audio), plus average and P90 latency by component
Transcript
The conversation, linked to recording playback; interrupted agent speech is marked when applicable
Extraction
Each configured extraction field with its extracted value, reasoning, allowed values, and prompt
Execution logs
An Execution Logs table (timestamp, log data, direction, component, provider, latency) and a Webhooks tab with webhook delivery details
Raw logs
The complete execution JSON, with Search logs, Copy JSON, and Download JSON. See the execution payload reference for what each field means
When you fetch logs via the Get execution raw logs API, LLM assistant responses may also include reasoning_content (model thinking or reasoning summary) when the provider returns it.
The raw logs format matches the Get Execution API response, making it easy to integrate with your systems programmatically. For a field-by-field breakdown of the JSON, see the execution payload reference.
Execution logs are essential for debugging. Use them to identify latency issues, transcription errors, or unexpected LLM responses.
