Privacy and data retention

Use of Aperture is governed by the Aperture Beta Terms.

Aperture captures the full request and response body, headers, and tool use data for every LLM request, alongside usage metrics such as token counts, model, and estimated cost. This hub gathers the controls for where that captured data is stored, how long it persists, and where request content leaves Aperture. Usage metrics are preserved regardless of your retention settings, so reporting and analytics remain available after Aperture purges request bodies.

Hosting location

Aperture instances are currently hosted in the European Union. Support for additional locations is planned for the future.

Control retention

Configure how long Aperture keeps captured request and response bodies before it purges them, or enable zero data retention so bodies are never written to disk. The retention fields (duration, purge, and require_export) are documented in the database configuration reference. For full zero data retention, also leave save_raws and keep_encrypted_blobs set to false, so Aperture keeps no raw or encrypted copies of request content.

Export and archive usage data

Export captured usage data to S3-compatible storage for long-term retention and archiving. You can require export before Aperture purges local captures, so no records are lost when retention expires.

Control third-party data egress

Control where request content leaves Aperture. Guardrails inspect, modify, or block requests before they reach the provider. Connectors centralize authentication and proxy outbound calls to external MCP servers and HTTP APIs. Integrations forward request and usage data to external enforcement and observability tools.