what sign up
nixm nixm

← learn

Is Claude private? What Anthropic keeps, and for how long

September 24, 2026

Private from other people, yes. Unstored, no. On the consumer plans (Free, Pro and Max) your chats stay in your account until you delete them, and a single setting decides the rest. Per Anthropic, with model improvement turned on, your chats can be used for training and kept in de-identified form for up to five years. With it off, deleted chats leave Anthropic's systems within 30 days and are not trained on. Incognito chats are never used for training. If what you want is a conversation that is never written down at all, that is what nixm is for.

The short answer

Claude is not public, it is not shown to other users, and Anthropic says it does not sell user data. That is one meaning of private. The other meaning is what happens to the words after you type them, and there Claude works like most assistants with an account: your conversations are stored so you can come back to them, and what happens next depends on your settings. Anthropic's own privacy centre lays it out clearly, and the answer turns on one switch.

The setting that decides it

In August 2025 Anthropic changed its consumer terms. Users of Claude Free, Pro and Max were asked to choose whether their chats and coding sessions could be used to improve Claude, and new users make the same choice at sign-up. The setting lives in Claude's privacy settings and can be changed at any time.

If model improvement is on, Anthropic says it may use your new or resumed chats for training and may keep them in a de-identified format for up to five years in its training pipelines. The five-year window only applies to chats started or resumed after the setting was turned on.

If model improvement is off, your chats are not used for training, and the standard 30-day back-end window applies to anything you delete.

Anthropic presents this as a choice. Several 2026 reviews describe the toggle as arriving switched on, so the safe move is to open the setting and look rather than assume which side you are on.

What deleting a chat actually does

Per Anthropic, a deleted conversation is removed from your chat history immediately and deleted from its back-end storage within 30 days. A deleted chat is also not used to train future models. Two things fall outside that. If the setting was on while a chat was already being used, Anthropic says the data can still be part of training runs already in progress and of models already trained. Nothing is pulled back out of a finished model. And turning the setting off stops future use of both your old and new chats, but it does not reach backwards into what has already been trained.

Incognito chats

Claude has an incognito mode. Anthropic says incognito chats are not used to improve Claude even if you have model improvement switched on. That makes incognito the clean option inside Claude for a conversation you do not want in the training pipeline. It is a promise about training, though. Like every chat, it still passes through Anthropic's systems to be answered.

The exceptions that run longer

Three kinds of data have their own clocks, per Anthropic's privacy centre:

  • Flagged chats. If automated trust-and-safety systems flag a conversation as violating the usage policy, the inputs and outputs can be kept for up to two years, and the classification scores for up to seven.
  • Feedback. A thumbs-up, thumbs-down or bug report keeps the data attached to that submission for five years.
  • Legal holds. Chats can be kept as required by law or to resolve disputes, in every case.

None of this is unusual for the category. It is simply the part people skip when they read "30 days".

Work accounts are a different product

Everything above is about the consumer plans. Claude for Work — Team and Enterprise — and the API run under Anthropic's commercial terms, where Anthropic says it does not train on customer content by default. A company workspace also puts your chats under your employer's admin controls. Being private from the model is not the same as being private from your company, which is the same distinction drawn for ChatGPT in what enterprise admins can see.

So is Claude private?

With model improvement off and a habit of deleting what you do not need, Claude's consumer retention is among the shorter ones in the field: 30 days after deletion, no training. With it on, a de-identified copy can outlive your account's own history by years. That is a real choice, and it is worth making on purpose. How Claude compares with the others is laid out in what each AI chatbot keeps, and the four different meanings of "private" are in the most private AI chatbot.

The version with nothing to delete

Every setting on this page is about managing a record after it exists. nixm starts from the other end: the conversation is never written to a database, so there is no history to delete and nothing for a training pipeline to keep. What remains of a session is a small signal that dissolves after seven days. It is not a general-purpose assistant and does not try to replace one. It is a place to think out loud without leaving a trail. How ephemeral AI works.

frequently asked

Does Claude save your conversations?

Yes. On the consumer plans your chats stay in your account until you delete them. Per Anthropic, deleted chats disappear from your history immediately and from its back-end systems within 30 days.

Does Claude train on my chats?

Only if the model improvement setting is on. Anthropic says it then may use new or resumed chats for training and keep them de-identified for up to five years. With the setting off, your chats are not used for training.

How long does Anthropic keep Claude chats?

With model improvement off, deleted chats are removed from back-end storage within 30 days. With it on, a de-identified copy can stay in training pipelines for up to five years. Flagged chats can be kept up to two years, and feedback for five.

Are Claude incognito chats private?

Anthropic says incognito chats are never used to improve Claude, even with model improvement switched on. They still pass through its systems to be answered.

If I turn the setting off, is my old data removed from training?

No. Anthropic says turning it off stops future training on both old and new chats, but data already in training runs or in trained models stays there.

Is Claude for Work more private than the consumer app?

On training, yes — Anthropic says it does not train on commercial customers' content by default. But a work account sits under your employer's admin controls, which is a different kind of visibility.

Can I try nixm without an account?

Yes. Open nixm.ai and seed a signal — no account needed, and the conversation is never written to a database.

try it now — no account, nothing stored. even the signal dissolves in seven days.

seed a signal →

Enter your email and receive a link to sign in. No passwords, just a secure signal to connect.

why connect

  • 20 free credits to start
  • your signals, remembered
  • buy more credits anytime
  • image & video generation features
or
+