Claude: what Anthropic says about your data
On Free, Pro and Max, each user chooses whether chats may be used for model training, and that choice changes how long chats are kept. Team, Enterprise and the API are under separate commercial terms.
Sources checked November 7, 2026
Personal plans
Claude Free, Pro and Max
- Training on your content
- Anthropic’s Aug 28, 2025 update says users choose a model training setting; chats are used for training only when it is on.
- How long it is kept
- Five years if the user allows training; the existing 30-day period if not.
- Your controls
- The model training setting in the user’s privacy settings. Each person sets it for their own account.
Business plans
Claude for Work (Team and Enterprise), the API, Amazon Bedrock, Google Cloud Vertex
- Training on your content
- The update says it does not apply to services under Anthropic’s Commercial Terms.
- How long it is kept
- Not covered by this page; see Anthropic’s Commercial Terms.
- Your controls
- Not covered by this page.
In Anthropic's words
What to write in your AI register
- Which plan your business uses, and whether staff sign in with company or personal accounts.
- Whether the training setting is on or off, who checked it and when.
- What kinds of data staff may put in (for example: no client names, no payment data).
Sources
Policies change; check the linked page before you decide. This page describes what the vendor says. It is not legal advice and not a review of the vendor.