← Starlog

AI for Relationships

Does Claude Have Memory? How to Turn It On and Use It

Generate memory from chats lets Claude carry context between chats. Where the setting is, what it saves and skips, and how it differs from projects.

Updated October 20, 2026 Intriq Editorial 11 min read
AI for Relationshipsaiassistantbriefing
Illustration of a figure in a portrait frame with four small person cards all pointing back at it, and one empty card standing apart

Yes. Claude has memory, and the switch is called Generate memory from chats. It lives in Settings > Memory, and with it on, Claude saves useful context as you talk and carries it into your next conversations. Anthropic says memory is on by default for Free, Pro and Max on the web, Claude Desktop and Claude Mobile, while on Team and Enterprise plans it stays off for each member until they turn it on, and only if an owner has made it available.

If you searched for “allow Claude to generate memory from your chats”, that is the setting you are looking at. The rest of this page covers what it actually saves, what it skips on purpose, and how it differs from chat search and projects, which are separate features that are easy to mistake for it. Every step below was checked against Anthropic’s memory help article on 20 October 2026.

Where the setting lives

  1. Open Settings.
  2. Select Memory.
  3. Turn on Generate memory from chats.

The same page holds everything else that matters: the list of what Claude remembers, the Search and reference chats toggle, Include sensitive topics in memory, and Start import for bringing memory over from another assistant.

If you see Memory under Settings > Capabilities instead, your account is on the legacy experience. Anthropic says a small number of Team and Enterprise organizations still use it. The older name you may have seen, Generate memory from chat history, is how the help article labels the organization-wide toggle in that legacy setup.

If the setting is missing on a work account, your owner controls it. Owners and Primary Owners turn memory on under Organization settings > Capabilities, and after that each member manages their own. Anthropic notes that turning memory off at the organization level permanently deletes every member’s memory, and that memory is not available to organizations with HIPAA, public-sector or custom data retention agreements.

What generating memory from chats actually does

The current version saves memory as a set of individual topics while you chat, rather than summarizing a conversation after it ends. 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 save something directly.

Anthropic lists the kind of context it keeps:

  • your role, projects and professional context
  • the people and places in your work and life
  • communication preferences and working style
  • technical preferences and coding style
  • project details and ongoing work

Memory is also shared with Claude Cowork when Cowork runs in the cloud, so a task you hand off already knows who your manager is and how they like updates written. Cowork sessions that run locally on your computer do not use memory.

What it does not save

This is the half most explanations leave out, and it is where expectations go wrong.

  • Sensitive topics are off by default. Claude does not store things like your health, race, ethnicity, religious beliefs, politics or gender identity unless you turn on Include sensitive topics in memory. Turning it on applies going forward only, and nothing from before is saved retroactively.
  • Some things are never saved, even if you ask: government ID numbers, criminal history, financial account numbers and immigration status.
  • Incognito chats are not saved to memory or to your chat history. Start one from the ghost icon in the upper right of a new chat outside a project.
  • Chats started with memory off neither use nor add to memory, though they stay in your history.
  • Pausing stops the clock. Conversations you have while memory is paused are not added later when you unpause.

One more detail that surprises people: deleting a conversation does not remove the memory generated from it. If a chat produced a topic you want gone, delete the topic itself.

Reading and editing what Claude remembers

Everything is listed under Topics in Settings > Memory. Select a topic to read it, use the edit icon to change it, or select Delete to remove it. A fix in one topic applies to every conversation from then on. You can also tell Claude in any chat what to remember, change or forget.

Open the list every few weeks. A wrong entry costs more than a missing one, because it gets applied quietly to every answer after it. The usual way it goes wrong is generalisation: you ask for something terse once on a rushed day, Claude saves a preference for terse writing, and three weeks later everything arrives clipped. You start adding “in full prose” to your prompts, when the real fix is ten seconds in the Topics list.

HOW ONE ENTRY COMPOUNDSOne terse request, on one busy daysaved as a standing preferenceEvery answer after it arrives clipped.
A single entry keeps applying itself until you go and read the list.

Turning it off: pause, reset, or one chat only

Switch off Generate memory from chats and Claude offers two choices:

OptionWhat happensReversible?
Pause memoryClaude keeps its existing memory but stops using it and stops making new memoriesYes. Unpause and it comes back, minus anything said while paused
Reset memoryPermanently deletes all memories, including project memoriesNo. You start from scratch
Memory off for one chatClaude neither uses nor saves memory in that chat; the chat stays in your historySet before the first message, then locked

To turn memory off for a single chat on the web or Claude Desktop: start a new chat, click + in the message box, click Memory to turn it off, then send your first message. A crossed-out memory icon then appears next to the chat title.

On Claude Mobile: start a new chat, tap + next to the message box, turn off the Memory switch, close the sheet, and send your first message. The switch stays visible but locked in the + menu.

The option only appears when memory is on for your account, and it does not appear in incognito chats.

Memory vs chat search vs projects

Three features carry context between conversations in Claude, and they work differently enough that it pays to know which one you are relying on.

MemorySearch past chatsProjects
What it isTopics Claude saves about you as you chatClaude looks up earlier conversations when you askA workspace with its own chats, files and instructions
How it kicks inAutomatically, in every chat with memory onWhen you ask, shown as a tool call with citationsWhen you chat inside the project
PlansFree, Pro, Max; Team and Enterprise when enabledPro, Max, Team, EnterpriseAll plans (free accounts get five projects)
SettingGenerate memory from chatsSearch and reference chatsPer project, including Use account memory
ReachYour account, plus each project’s own memoryChats outside projects, or within one projectThat project only

Chat search retrieves the conversation itself. Ask “what did we discuss about the pricing page?” and Claude searches your history, cites the chats it found, and offers to delete them. Memory, by contrast, is the distilled version that is already present before you ask anything.

Projects each get their own memory. Inside a project, Claude uses the project’s memory alongside your account memory from chats outside projects, but never another project’s. To keep a project fully separate, turn off Use account memory under Memory on the project’s page. Projects also hold a knowledge base of uploaded files and project instructions, which are context you write deliberately rather than context Claude infers.

When Claude does not seem to remember anything

It is usually a switch, not a fault. Work through these in order:

  1. Memory is off. On Team and Enterprise it starts off for every member, and your owner has to make it available first.
  2. Memory is paused. Everything from before is intact, and nothing since has been saved.
  3. That chat started with memory off, or it was an incognito chat.
  4. The detail was a sensitive topic and Include sensitive topics in memory is off.
  5. You are in a project with Use account memory turned off.
  6. You are on the legacy experience, with memory under Settings > Capabilities rather than Settings > Memory.

If none of those apply, try asking it to search your past chats instead. On paid plans that reaches the full conversation, not just what memory chose to keep.

Importing memory from ChatGPT or Gemini

Anthropic’s import and export article gives the steps, which work on the web and Claude Desktop for every plan:

  1. Go to Settings > Memory and select Start import.
  2. Copy the prompt the import flow shows you and run it in the assistant you are leaving.
  3. Paste what it returns into the text box and click Add to memory.
  4. Click See what Claude learned about you to check the result.

Anthropic is candid that imports are experimental and may not always land, and that because memory focuses on work-related topics, Claude may not retain imported personal details unrelated to work. To export in the other direction, ask Claude in a chat to write out its memories of you verbatim and save the result in a file.

WHAT YOU TALKED ABOUTPriya's mother's scanthe intro you promisedwhy the deal stalledMarcus defends in Julywho to avoid at AcmeWHAT ARRIVESTone and formatRole and projects
The import carries a profile, not an archive.

What this means for the people in your life

Claude’s memory does now keep “the people and places in your work and life”, which is a real step forward from a memory that only described the account holder. It knows who your manager is. It can know which client prefers a call to an email.

The shape of what it keeps still follows from what it is for. It focuses on work. It decides on its own what is worth saving rather than keeping what you said. And it leaves sensitive topics out unless you opt in.

Say you had this conversation in June:

Call with Priya, 3 June. She has left the agency and is consulting solo, underpricing herself badly. Her mother’s scan came back clear. Wants an intro to someone who has run a two-sided marketplace. Said to check in after Lisbon in August.

Paste that into Claude and you will get a warm, well-judged follow-up message straight away. Whether any of it is waiting for you in November depends on what memory chose to keep. The detail about her mother is a health topic, so by default it is not stored at all, and that is the line you would most regret getting wrong when you next see her.

That is a sensible default for an assistant. It is also why a dated note about a person belongs in a place built to keep exactly what you wrote. A relationship memory app files it under Priya, keeps the date, and answers from your own words rather than filling the gap with something plausible.

“What did Tariq need?”GROUNDEDIntros to operators,not investors.FROM YOUR NOTE, 14 MARINVENTEDSeries A intros anda warm VC list.NOT IN YOUR NOTES
The useful answer names its source. The dangerous one sounds identical.

For how the other assistants handle this, see what ChatGPT, Claude, Gemini and DeepSeek remember, and if you are coming from ChatGPT, what to save before clearing ChatGPT memory.

Key takeaway: Generate memory from chats is the setting in Settings > Memory that lets Claude save context as topics while you chat. It is on by default for Free, Pro and Max, and off for Team and Enterprise members until enabled. It skips incognito chats, chats started with memory off, and sensitive topics unless you opt in, and deleting a chat does not delete its memory. Chat search and projects are separate features with their own settings.

FAQ

What does “Generate memory from chats” do in Claude?

It lets Claude save useful context from your conversations as individual topics, such as your role, ongoing projects, preferences, and the people and places in your work, and use them in later chats. You can read, edit and delete every topic in Settings > Memory.

How do I turn on memory in Claude?

Open Settings, select Memory, and turn on Generate memory from chats. On Free, Pro and Max it is on by default. On Team and Enterprise an owner has to make memory available first, and then each member turns it on for themselves.

Does free Claude have memory?

Yes. Anthropic says memory is on by default for Free, Pro and Max on the web, Claude Desktop and Claude Mobile. Searching past chats is the part that needs a paid plan.

Does Claude have memory across chats?

Yes. With memory on, what Claude saves in one chat is available in your next ones. Inside a project, Claude uses the project’s own memory plus your account memory, but not memory from other projects, unless you turn off Use account memory for that project.

Do Claude projects have memory?

Yes. Each project has its own memory, alongside its uploaded files and instructions. New memory from a project chat is saved to that project, and if you have memory on, Claude can reference it in chats outside projects too.

What is the difference between pausing and resetting Claude’s memory?

Pausing keeps everything Claude remembers but stops it using or adding to memory, and anything said while paused is never added later. Resetting permanently deletes all memories, including project memories, and cannot be undone.

Does deleting a chat delete Claude’s memory of it?

No. Anthropic says memory generated from a conversation is not removed when that conversation is deleted or expires. Delete the specific topic in Settings > Memory instead.

Can Claude import my ChatGPT memory?

Yes. Go to Settings > Memory, select Start import, run the prompt it shows in ChatGPT, paste the result, and click Add to memory. Imports are experimental, and personal details unrelated to work may not be kept.

Intriq files dated notes under each person and answers only from what you wrote. Start with the personal CRM hub, read can AI remember people for you, or see how AI memory compares to human memory for relationships.

Meet Mira.

Scan with your phone or choose your store.

App Store Google Play

Free to start.