Zero Data Retention
For the governing documents, see Anam’s Privacy Policy, AI Governance Practices, and Data Processing Addendum. Review these alongside your agreement and deployment configuration rather than treating an integration mode as a retention policy. Enterprise customers can enable Zero Data Retention at a persona level or at the session level (for ephemeral sessions). When enabled, Anam does not persist any customer content - including conversation transcripts, user audio, TTS text, LLM prompts/responses, or session recordings. Processing occurs in volatile memory only. Anam retains operational metadata necessary for billing, service delivery, and security, including: session timestamps and duration, usage metering, error codes, performance metrics, and infrastructure logs (retained for 30 days). Account-level data (organization info, API keys, subscription) is retained as standard. Support and troubleshooting capabilities may be limited for ZDR-enabled sessions. Zero Data Retention does not apply to any images used in one shot persona creation and/or any files uploaded to a persona’s knowledge base.Enabling Zero Data Retention
On Enterprise plans, Zero Data Retention can be enabled via theAdvanced tab of the lab, or via persona configuration for either stateful or ephemeral personas. See the API Reference for more details.
Not all LLMs and voices support zero data retention mode, when enabling zero data retention you may have to switch LLM or voice before you are able to publish and use your persona.
Confirm retention for your deployment
Under section 6 of Anam’s Privacy Policy, non-ZDR session recordings are retained for up to 30 days. Session reports, including transcripts, analytics, summaries, and generated insights, remain until the customer deletes them. A recording may be unavailable sooner. A signed download URL can also expire before the recording is deleted. Check the policy and your agreement for the applicable terms. For audio-passthrough and turnkey deployments, confirm which rules apply to each data category:- Session audio and video, including conversation recordings.
- Session reports, transcripts, analytics, and summaries.
- Operational metadata and logs.
- Source images used to create avatars and uploaded knowledge files.
- Files produced by script-to-video jobs.
- Data sent to external LLM, TTS, and other integration providers.

