Model Rover

Data and Result Retention

Understand how request metadata, bodies, and generated media are retained.

The platform keeps request metadata needed for billing, usage analysis, reliability, and support. Metadata can include the model, API key identifier, status, timing, usage, and final amount.

Request and response bodies

Sanitized request and response bodies can be available in request logs for the retention period configured by the site. Binary, base64, and oversized values can be replaced with placeholders, and expired bodies cannot be restored from the log view.

Embedding vectors are returned to the caller but are not stored in request-log bodies.

Generated media

Image and video URLs can be temporary. Some results are transferred to platform storage, while others remain provider URLs; neither should be treated as permanent application storage. Download and copy assets you need to retain.

Sensitive data

Send only data you are authorized to process. Do not place API keys, passwords, payment credentials, or unnecessary personal data in prompts, uploaded files, callback URLs, or metadata.

Help us improve this page

Found something unclear, outdated, or incorrect?

Last updated on