Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

Zero data retention

Zero data retention (ZDR) is an arrangement in which an AI provider does not keep a customer's prompts and outputs after a request is answered, usually offered only to approved customers and only for eligible API features.

Also called: ZDR

By Divya Sharma · Checked Nov 15, 2026 · How we review

By default, hosted AI APIs keep some of what you send. OpenAI, for example, says abuse monitoring logs, which can include prompts and responses, are kept for up to 30 days. Zero data retention removes that storage for customers who qualify, so the provider handles each request and then keeps no copy of the content.

How it works in practice

ZDR is a contract and account setting, not an API flag you switch on yourself. OpenAI requires prior approval and offers it alongside a lighter option called Modified Abuse Monitoring. Anthropic lists, feature by feature, which parts of its API are eligible. Features that need storage to work, such as code execution containers, are not covered or keep a small technical record for a short time. Anthropic also notes that some models require 30-day retention and that third-party integrations fall outside the arrangement.

How it differs from related terms

ZDR is about storage, not training. Many providers already say API data is not used to train models; see AI training data for how that varies. It is also not data residency, which controls where data is processed rather than whether it is kept. Features such as prompt caching may still hold short-lived data in memory. If sensitive data must never leave your systems, PII redaction before the call or self-hosting are the stronger options. Check your own contract for what applies; this page describes the term, not your obligations.

Related terms

Data residency
Data residency is a control that keeps a customer's data stored, and often processed, in a chosen geographic region such as the EU, so a team can meet its own legal, contractual or customer rules on where data lives.
AI training data
AI training data is the collection of examples a model learns from, such as web text, code, images, transcripts or labelled pairs. Its size, quality, mix and licensing shape what the model can do and where it fails.
PII redaction
PII redaction is the detection and removal or masking of personally identifiable information, such as names, phone numbers, addresses and card numbers, from text or images before it is sent to a model, stored in logs or shown to others.
Self-hosting (AI models)
Self-hosting an AI model means running it on hardware you control, such as your own servers, rented GPUs or a laptop, instead of calling a provider's API. You choose the model and keep the data, but you run the infrastructure.
SOC 2
SOC 2 is an audit report, produced by an outside CPA audit firm under AICPA standards, on a service organisation's controls relevant to security, availability, processing integrity, confidentiality or privacy.

Sources