Claude Memory: How It Works and How to Control It

Claude saves memory as individual topics while you chat, keeps each project's memory separate, and gives you a full view of everything it holds. Here is how it actually behaves — and every control you have over it.

Last updated August 2026

Heads up: some links below are referral links. They cost you nothing extra and never change what we recommend. How this site makes money.

The short answer

Claude's memory lets it carry context between conversations — your role, your projects, your preferences — so you stop re-explaining yourself in every new chat. It is on by default for Free, Pro and Max plans on the web, desktop and mobile apps. On Team and Enterprise plans it is off by default until an organisation owner enables it.

Two things people most want to know up front: you can see and edit everything it stores in Settings → Memory, and sensitive personal topics are excluded by default. The details, verified against Anthropic's own support documentation in September 2026, are below.

How Claude's memory actually works

Claude saves memory as a set of individual topics as you chat, rather than summarising conversations after they end. Mention that a deadline moved and the next conversation already knows. Claude decides what to save on its own, and you can also say “remember this” to store something directly.

Project memory is separate

Each project has its own memory space and its own project summary. Context inside a project stays there — it does not leak into other projects or into your ordinary chats, and vice versa.

One memory across chat and Cowork

What Claude remembers from your chats is available when you hand it a task in Claude Cowork running in the cloud, and what comes up in a Cowork task carries back to chat. The caveat: Cowork sessions that run locally on your computer don't use memory — only cloud sessions share it.

Claude has two separate mechanisms for using your history, and they have different plan requirements.

Memory is the stored-topics system above — available on Free, Pro and Max.

Search past chats lets Claude look back through your previous conversations on request (“what did we discuss about X?”) and cite what it finds. This one is paid plans only — Pro, Max, Team and Enterprise. Searches show up as visible tool calls in the chat, with citations linking back to the original conversations. Search covers all chats outside projects; inside a project, it searches only that project's conversations.

If searching your history matters to you and you're on the free tier, that's one of the quieter differences the free vs Pro comparison comes down to.

What Claude remembers — and what it refuses to

Memory focuses on everyday working context: your role, projects and professional context, the people and places in your work, communication preferences and working style, technical and coding preferences, and ongoing work.

Sensitive topics are off by default

By default Claude does not store topics like your health, race, ethnicity, religious beliefs, politics or gender identity. If you'd rather not keep re-explaining one of these, you can opt in via Include sensitive topics in memory in Settings → Memory — and each time Claude then saves something sensitive, a notice appears so you can review it. Turn the setting off later and already-saved sensitive items are removed.

Some things are never saved, even if you ask

Government ID numbers, criminal history, financial account numbers and immigration status are never written to memory. Claude tells you when it can't save something for this reason.

Incognito chats leave no trace in memory

Click the ghost icon on a new chat (available on every plan, including free) and the conversation is excluded from both memory and chat history. One honest caveat for workplace accounts: on Team and Enterprise plans, incognito chats still appear in organisational data exports and follow your company's retention policies — incognito hides a chat from your history, not from your admin.

Viewing, editing, pausing and resetting

Everything Claude remembers is listed under Topics in Settings → Memory. Open any topic to read it, edit it, or delete it — a fix made there applies to every conversation from then on. You can also just tell Claude in a chat what to remember, change or forget, and the update applies to your next conversation.

Turning memory off offers two distinct options, and the difference matters:

One retention detail worth knowing: deleting a conversation does not delete the memory entries that were generated from it. If you want something gone from memory, delete the memory entry itself.

Moving memory in and out of Claude

You can import memory from another AI assistant: export your stored context from the other service, then paste it into the import flow at Settings → Memory → Start import. Claude extracts the key information into individual memory entries. Anthropic labels this feature experimental, and notes that imported personal details unrelated to work may not be retained — memory is deliberately work-focused.

Export works too: ask Claude to write out its memories of you verbatim, or view them in settings, and save the result as a backup or to carry to another service.

The Team and Enterprise picture

Memory is off by default for organisations, and only an owner can switch it on. Once enabled, individual members manage their own memory; owners cannot view or edit anyone's individual memories, and enabling memory does not enable sensitive topics — each user opts into those personally.

Two hard edges: if an owner turns memory off for the organisation, all existing memory entries for all users are deleted immediately and permanently. And memory isn't available at all to organisations with HIPAA, public-sector or custom data-retention agreements.

Common questions

Is Claude's memory available on the free plan?
Yes — memory is on by default for Free, Pro and Max. Searching past chats is the paid-only feature: it needs Pro, Max, Team or Enterprise.
How do I see what Claude remembers about me?
Settings → Memory lists everything under Topics. You can read, edit or delete any entry there, or ask Claude in a chat to change or forget something.
Does deleting a conversation delete its memories?
No. Memory entries generated from a deleted conversation remain until you delete them individually in Settings → Memory.
Can Claude remember things across projects?
No — each project has its own separate memory space and summary. Context in one project doesn't carry into another or into non-project chats.
What's the difference between pausing and resetting memory?
Pause keeps existing memories but stops using or adding to them; you can resume later. Reset permanently deletes everything, including project memories, and cannot be undone.

Memory works on the free plan

You don't need a subscription to try it — memory is on by default for free accounts. Searching past chats is the part that needs Pro.


Related