Call Recordings

Workspace-scoped call recordings with separate participant channels, synchronized playback, metadata, and waveform support.

When call recording is enabled, the voice runtime can retain the caller and agent audio as separate channels. Separate channels support participant-specific review while allowing the playback API to construct a synchronized stereo stream.

Recording finalization is best-effort. A call can complete successfully even when no recording artifact is available, so clients should check the call's recording-availability field rather than infer availability from call status.

Channels and Timing

Channel	Contents	Playback behavior
Caller	Inbound caller audio	Served as mono WAV for browser playback
Agent	Spoken agent audio	Served as mono WAV using the sample rate recorded in metadata
Stereo	Caller on the left, agent on the right	Constructed on request from the two stored channels

The two source channels can use different audio formats and sample rates. Recording metadata preserves the actual format, rate, duration, size, direction, and media-start timestamp for the call. The playback service converts or resamples audio when needed; the stored source channels are not assumed to be a prebuilt stereo file.

The runtime aligns outbound audio with the inbound timeline during capture. This shared timing origin lets playback views align transcripts, agent speech, tool activity, and other call events without treating each channel as an independent recording.

Playback and Review

Depending on the API surface used, authenticated clients can retrieve:

• Caller-only and agent-only audio
• A synchronized stereo WAV stream
• Recording metadata
• Aligned waveform peaks for timeline rendering
• Raw channel downloads for supported recording endpoints

Recording endpoints authenticate the requester and resolve the call within the requester's workspace before returning an artifact. Missing calls, unavailable recordings, and unavailable recording services return explicit errors rather than an empty audio response.

Availability and Retention

Recording availability depends on deployment configuration and successful finalization. Retention requirements should be configured and validated as part of the workspace's data-handling policy; the current runtime does not expose the former multi-tier archive-and-restore workflow described in older documentation.

For optional post-call transcript verification and quality analysis, see After the Call. For timeline review that aligns speech with agent actions, tool calls, and system events, see Call Playback Timeline.