Data retention

Phonic supports three data retention modes, configurable per workspace. They trade off operational features against how long your data lives on Phonic’s servers and how much of it reaches third-party providers.

All modes are available at no additional cost. Contact your account team to change the mode for your workspace.

Standard

The default for most workspaces. Transcripts and audio recordings are retained indefinitely so you can review past conversations, run evals after the fact, and debug agent behavior over time.

What we store:

• Audio recordings: full call audio in object storage.
• Transcripts: both raw and retranscribed text, per turn.
• Derived artifacts: conversation evals, data extractions, tool call logs, API request/response logs.

What we don’t store:

• Audio buffers older than the call. The streaming buffer is discarded once the call ends; only the finalized recording is persisted.
• Customer secrets passed in tool call payloads (these are redacted before logging).

Short retention

For workspaces that need conversation history operationally but want a bounded exposure window. Transcripts and audio are retained for 24 hours, then hard-deleted by a sweeper. Derived artifacts (evals, extractions) follow the same TTL.

Webhooks (conversation.ended, conversation.analysis, etc.) fire normally. If your async pipeline needs the data beyond 24h, fetch and store it on your own side before the window closes.

The retention window cannot currently be customized; it is a fixed 24 hours.

Zero retention (ZDR)

For workspaces with the strictest privacy requirements. Conversation content is never persisted on Phonic’s servers beyond the call itself. The streaming buffer exists only in memory during the call.

What’s retained under ZDR:

• Conversation metadata only: conversation_id, external_id, start/end timestamps, duration, turn count, agent ID, and routing metadata (transfer, hangup reasons).
• Billing aggregates (call minutes, model usage).

What’s never retained:

• Audio recordings.
• Transcripts (raw or retranscribed).
• LLM prompts and completions.
• Tool call inputs and outputs.

What breaks under ZDR

ZDR disables features that depend on stored content:

Webhooks still fire. conversation.ended carries the full final transcript in its payload, since that payload is the only opportunity for your systems to receive it. Once the webhook is delivered, the content is discarded.

Third-party provider behavior

Under ZDR, Phonic passes no-retention flags to all upstream providers:

• OpenAI: requests are sent with store: false. Phonic operates under OpenAI’s Zero Data Retention agreement, so prompts and completions are not retained by OpenAI for abuse monitoring.
• Deepgram: speech-to-text requests use the no-storage configuration.
• LiveKit / Twilio: call recording is disabled at the transport layer.

A current list of which providers and which models offer ZDR-compatible tiers is maintained on your workspace settings page.

Hard deletion on request

Regardless of retention mode, you can request hard deletion of specific conversations or all data tied to a specific end user. Hard deletion removes transcripts, audio, derived artifacts, and observability records within 24 hours.

1
DELETE /v1/conversations/{conversation_id}
2
DELETE /v1/users/{external_user_id}/data

These endpoints return 202 Accepted and complete asynchronously. A webhook fires when deletion completes.

Choosing a mode

• Most workspaces should use Standard. The default is appropriate for iterating on agent prompts and reviewing conversations.
• Use Short retention if your compliance program requires a bounded retention window but you still want conversation history available during a short review period.
• Use ZDR only if you’ve audited the feature trade-offs above and your security team requires zero persistence. ZDR is not a default; it is a deliberate trade.

For help selecting the right mode for your use case, contact your account team.