Claude Memory: How It Works, and How to View, Edit or Turn It Off (2026)
Claude memory is a set of short topics Claude saves from your chats, such as your role, projects and preferences, and carries into new conversations. It is on by default for Free, Pro and Max users and off by default on Team and Enterprise. You can view, edit and delete it in Settings, Memory, pause it, or reset it. Deleting a chat does not delete the memories made from it (Anthropic Help Center, 2026).
This guide is based on Anthropic's two Help Center articles on memory, "Use Claude's chat search and memory to build on previous context" and "Importing and exporting your memory from Claude", both read in September 2026. It explains what memory keeps, how it differs from Claude's past-chat search, the two ways to turn it off, how to move memory to or from ChatGPT, and when you need a searchable record of the full conversations instead, which is where AI Toolbox (formerly Claude Toolbox), a Chrome extension with 40,000+ users and a 4.7/5 Chrome Web Store rating, fits in.
By Adi Leviim, Co-Founder at Infi Developments. 7+ years building AI productivity tools. Last reviewed September 2026.
What Claude Memory Is, and What It Is Not
Claude has two separate features that both look like "memory", and they work differently. Memory saves short topics about you as you chat. Chat search lets Claude look through your past conversations when you ask it to. Anthropic describes memory as Claude saving "a set of individual topics as you chat, rather than summarizing conversations after they end" (Anthropic Help Center).
Feature
What it does
Plans
Where to control it
Memory
Saves topics such as your role, projects, people and preferences, and uses them in new chats
On by default on Free, Pro and Max; off by default on Team and Enterprise until an owner turns it on
Settings, Memory, "Generate memory from chats"
Search past chats
Searches earlier conversations when you ask, shown as a tool call with citations to the original chats
Pro, Max, Team and Enterprise
Settings, Memory, "Search and reference chats"
Project memory
Each project keeps its own memory and summary, separate from other projects and from non-project chats
Where projects are available
Inside each project
Incognito chat
A temporary chat that is not saved to history or memory, and never appears in chat search
All plans
Ghost icon at the top right of a new chat, outside projects
Memory is available on the web, Claude Desktop and the mobile apps. It is also shared with Claude Cowork when Cowork runs in the cloud; Cowork sessions that run locally on your computer do not use memory.
What Claude Remembers, and What It Never Saves
Claude memory is built around work context. According to Anthropic, it keeps things like:
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
By default Claude does not store sensitive topics such as health, race, ethnicity, religious beliefs, politics or gender identity. You can opt in with Include sensitive topics in memory; it only applies from that point forward, and turning it off again removes the sensitive items already saved. Some information is never saved even if you ask: government ID numbers, criminal history, financial account numbers and immigration status.
How to See and Edit What Claude Remembers
Everything Claude remembers is listed in one place.
Open Settings in claude.ai, Claude Desktop or the mobile app.
Select Memory. Your memories appear under Topics.
Select a topic to read it, use the edit icon to change it, or select Delete to remove it. A fix applies to every conversation from then on.
You can also manage memory from inside a chat: tell Claude what to remember, change or forget, and the update applies to your next conversation. To force something in, say "remember this". If you see Memory under Settings, Capabilities instead, your account is still on Anthropic's older memory experience, which a small number of Team and Enterprise organizations use.
How to Turn Off Claude Memory: Pause or Reset
Turning off Generate memory from chats gives you two choices, and only one of them deletes anything.
Option
Existing memories
New memories
Can you undo it?
Pause memory
Kept, including sensitive topics if you had them on
Not created, and chats during the pause are never added later
Yes, unpause to use memory again
Reset memory
Permanently deleted, including project memories
Start from scratch if you turn memory back on
No
To stop Claude from reading your old chats without touching memory, turn off Search and reference chats in the same page. For a single conversation you never want remembered, start it as an incognito chat.
Deleting a Claude Chat Does Not Delete Its Memories
This is the part most people get wrong. Anthropic states that when a conversation expires or is deleted, "related memory entries generated from it won't be removed". Deleting the chat removes the transcript, but the topics Claude learned from it stay until you delete them yourself in Settings, Memory.
To remove something completely: delete the chat, then delete the matching memory topic.
To keep a chat but drop what Claude learned: delete only the memory topic.
To erase everything Claude learned: use Reset memory.
Memory is also included in your data export, so an account export captures it alongside your conversations. The guide to deleted and missing Claude conversations covers what happens to the chats themselves.
Memory keeps a few facts. Your chats hold the whole conversation. AI Toolbox (formerly Claude Toolbox) adds full-text search across every Claude message, folders, bookmarks and export to TXT, Markdown, JSON or PDF, so the details memory leaves out stay easy to find. 40,000+ users, 4.7/5 on the Chrome Web Store.
Claude can import memory from another assistant and give you a copy of its own. Anthropic calls the feature experimental and says Claude "may not always successfully incorporate imported memories". Imports are available on Free, Pro, Max and Team, on the web and Claude Desktop (Anthropic Help Center).
Import into Claude:
In ChatGPT or another assistant, ask it to list every memory and piece of context it has about you in a single code block. Claude's import screen shows Anthropic's recommended prompt for this, which asks for each entry verbatim with its date.
In Claude, open Settings, Memory and select Start import.
Paste the exported text and select Add to memory. Claude splits it into individual topics you can review.
Remove anything sensitive before pasting. Because Claude's memory focuses on work, it may drop imported personal details that are unrelated to work.
Export from Claude: ask Claude, "Write out your memories of me verbatim, exactly as they appear in your memory," then save the answer to a file. That copy can be pasted into another assistant's memory. For ChatGPT's side of the move, the ChatGPT memory guide explains its settings.
Claude Memory on Team and Enterprise Plans
On Team and Enterprise, memory and sensitive topics are two separate organization settings, and both start off. Owners and Primary Owners turn memory on in Organization settings, Capabilities; each member then manages their own memory, and owners cannot view or edit a member's individual memories. Turning memory off for the organization deletes every member's memory immediately. Memory is not available to organizations with HIPAA, public-sector or custom data retention agreements, and search of past chats does not work when an Enterprise organization uses customer-managed encryption keys.
Memory vs a Searchable Record: Native Claude vs AI Toolbox
Memory is deliberately selective: it keeps topics, not transcripts. When you need the exact wording of an old answer, a code block or a decision from three months ago, you need the conversation itself. That is what AI Toolbox's Claude module adds.
Task
Claude (built in)
AI Toolbox Free
AI Toolbox Premium
Carry preferences into new chats
Memory
Not a memory feature
Not a memory feature
Find an exact phrase in an old chat
Title search; paid plans can ask Claude to search past chats
Full-text search of every message, first 5 results
All results, plus Advanced Search filters
Bring one specific past chat into a new one
Ask Claude to search for it
Type @@ and pick the chat, 3 times a day
Unlimited @@, several chats per message
Keep a copy outside Claude
Full account export
Export a chat as TXT
TXT, Markdown, JSON, PDF, and bulk export
Group chats by project or client
Projects
2 folders
Unlimited folders
@@ Context Mentions is the deliberate version of memory: you choose exactly which conversation Claude sees. Short chats are inserted as they are. Long ones are summarized first by AI Toolbox's summarizer, which sends the selected messages to an AI model through AI Toolbox's server and does not store them. The Claude context mentions guide shows it step by step, and the Claude search guide covers full-text search.
Who Should Use Claude Memory, and How
Freelancers and consultants with several clients: keep each client in its own project, so project memory stays separate, and use incognito for one-off questions you do not want remembered.
Developers: memory picks up coding style and tools; for exact code from an old session, search the transcript instead of relying on memory.
Anyone switching from ChatGPT: run the import once, then review the topics Claude created and delete anything wrong.
Privacy-minded users: leave sensitive topics off, check Settings, Memory every few weeks, and remember that deleting a chat does not clean its memories.
Limitations
Anthropic's memory is selective by design, so it will not recall every detail you mentioned, and import is still experimental. Memory also does not replace a backup: it is a handful of topics, not your conversations. AI Toolbox does not read or change Claude's memory; it works on the chats themselves. Plan details and setting names above come from Anthropic's Help Center as of September 2026 and can change.
Frequently Asked Questions
Does Claude have memory?
Yes. Claude saves short topics from your chats, such as your role, projects and preferences, and uses them in new conversations. Memory is on by default for Free, Pro and Max plans on the web, desktop and mobile apps, and off by default on Team and Enterprise until an owner enables it.
How do I see what Claude remembers about me?
Open Settings, then Memory. Everything Claude remembers is listed under Topics. Select a topic to read it, use the edit icon to change it, or select Delete to remove it. You can also tell Claude in a chat what to remember, change or forget.
How do I turn off Claude memory?
In Settings, Memory, turn off Generate memory from chats and choose Pause or Reset. Pause keeps existing memories but stops Claude using or adding to them. Reset permanently deletes all memories, including project memories, and cannot be undone.
Does deleting a Claude chat delete its memory?
No. Anthropic says memory entries generated from a conversation are not removed when that conversation is deleted or expires. Delete the chat, then delete the matching topic in Settings, Memory, or use Reset memory to clear everything Claude learned.
Can I import my ChatGPT memory into Claude?
Yes, as an experimental feature on Free, Pro, Max and Team plans on the web and Claude Desktop. Ask ChatGPT to list all its memories about you, then in Claude open Settings, Memory, select Start import, paste the text and select Add to memory.
Does Claude remember incognito chats?
No. Incognito chats, started from the ghost icon outside projects, are not saved to your chat history or to memory, and Claude never pulls them into past-chat search. They are available on every plan. On Team and Enterprise they still follow the organization's data retention and exports.
Is Claude memory the same as searching past chats?
No. Memory saves short topics and applies them automatically. Search past chats, available on paid plans, looks through your earlier conversations only when you ask and cites the original chats. You can turn each one on or off separately in Settings, Memory.
Bottom Line
Claude memory is a short list of topics, and you control it from Settings, Memory: edit or delete topics, pause to keep them, reset to erase them. The one trap is that deleting a chat does not delete what Claude learned from it. For the full conversations behind those topics, keep a searchable, exportable record.
A Full Stack Developer with 7+ years of experience building AI productivity tools. Leads product development and frontend architecture for AI Toolbox, the Chrome extension suite (ChatGPT, Gemini, Grok, and Claude modules) that helps users search, organize, and export their AI conversations.
Liked this guide? Get the next one first.
Join 40,000+ readers getting practical prompts, new features, and subscriber-only deals. No spam, unsubscribe in one click.
New features and product updates before anyone else
Time-saving prompts and workflows for ChatGPT, Gemini, Claude, and Grok
Adi Leviim