> ## Documentation Index
> Fetch the complete documentation index at: https://docs.primalabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Zero data retention

> What PrimaLabs stores about your requests — and what it never stores.

PrimaLabs runs the serverless inference API with **zero data retention** for request and
response content. Prompts and completions are processed in memory to serve the request and
are not written to logs, databases, or exports.

## What is never stored

* Prompt and message content, including system prompts.
* Model output, including streamed tokens.
* Tool-call names and arguments.

## What is stored

Billing and abuse prevention need request **metadata**, never content:

| Retained                             | Why                                         |
| ------------------------------------ | ------------------------------------------- |
| Model ID, timestamp, latency         | Usage reporting in the dashboard            |
| Token counts (input, output, cached) | Per-token billing                           |
| API key and account identifiers      | Attributing spend to your account           |
| HTTP status                          | Error rates, rate-limit and abuse detection |

## Scope

Zero data retention applies to every serverless endpoint under `https://api.primalabs.ai/v1`
and to every account — it is not a per-key setting and there is nothing to enable.

<Info>
  Questions about data handling for a dedicated deployment or a compliance review?
  [Contact us](https://primalabs.ai/book-demo).
</Info>