Ask ChatGPT what it knows about you and it will tell you something. What most people do not realise is that the screen where you manage that knowledge changed completely on 4 June 2026. OpenAI replaced the familiar scrolling list of saved memories with a single memory summary, written and rewritten automatically by a background process the company calls dreaming. The old list still exists, but it is now the legacy option.
That matters because many guides to ChatGPT memory still walk you through the old path, and following it on a current account leads nowhere. This guide covers how memory actually works now, the exact steps to delete it, what deleting does and does not remove, how memory behaves inside projects and Temporary Chats, and the one behaviour that catches people out: switching memory back on can bring the memories back.
The Key Takeaways
- New system since 4 June 2026: a continuously updated memory summary replaced the saved memories list, built by OpenAI's dreaming process.
- Where to look: Settings, then Personalization, then Memory. Not Data controls, which is where every stale guide sends you.
- Deleting is not erasing: clearing your memories does not delete your chats, and turning memory back on lets ChatGPT rebuild memories from the chats still in your history, including older ones.
- Muting is not deleting either: the newer Don't mention this again option reduces future references to a detail but leaves the original chat, file or email in place.
- Full removal takes several deletions: past chats, archived chats, files in Library or Space, the memory summary, and any connected apps holding the same detail.
- Screens still differ: OpenAI began with Plus and Pro in the US, and as of November 2026 its Memory help page still says controls vary by plan, region, platform and workspace, so two accounts can legitimately show two different screens.
How ChatGPT Memory Works Now
Memory is the context ChatGPT carries from one conversation into the next. When it is on, ChatGPT pulls useful details from your chats, your files and your connected apps, so you do not have to reintroduce yourself every time you open a new window.
The important shift is in how those details are produced. Under the old system, ChatGPT wrote a note whenever something looked worth keeping, and those notes piled up in a list you maintained by hand. Under the new one, a background process reads across your conversation history and keeps a synthesis of it up to date. That synthesis is one of the hidden inputs behind why two people get different answers to the same question. You read that synthesis on the memory summary page.
The memory summary
The memory summary is a written, high-level portrait of what ChatGPT has concluded about you. You can change it three ways: type a correction into the text field, highlight a line and correct that specific piece, or select Don't mention this again where the option appears.
OpenAI is direct about its limits. The summary does not necessarily include every detail or source that memory can reference, and it may leave out details that are less relevant or come from sources that are not shown individually. If you want to know whether something specific is in there, the company's own advice is to ask ChatGPT what it remembers about you.
Where the context actually comes from
When ChatGPT uses personal context, Sources can appear below the response. It lists what fed the answer: a past chat, a saved memory, a custom instruction, a file or a connected email. From there you can open the source, correct or delete a saved memory, delete a referenced chat, or mark a source as relevant or not relevant. Two caveats are worth carrying: Sources may not show every factor that shaped a response, and memory Sources are not included when you share a conversation through a link.
The inputs are wider than chats. Depending on your plan and region, OpenAI's Memory help page lists past chats, saved memories, custom instructions, files saved in Library or Space, and content from connected apps such as Gmail. Memory can also shape a web search: ChatGPT may use a detail such as your location or a dietary preference to write a better search query.
What Changed in June 2026, and Why
OpenAI published the reasoning in a post titled Dreaming: Better memory for a more helpful ChatGPT on 4 June 2026. Memory first launched in April 2024 as saved memories, a system the company now likens to talking to someone who took a few notes but still forgot everything that was not written down. In April 2025 it added the ability to reference chat history, powered by the first version of dreaming. The June 2026 release rebuilt the whole thing on that foundation.
The problem being solved was staleness. In OpenAI's words, saved memories "tend to go stale over time and eventually become incorrect or irrelevant". The company's own test case is a note that you are planning a birthday party for next Saturday: eventually, Sunday arrives. A synthesis that updates itself handles the passage of time, so "You're going to Singapore in July" becomes "You went to Singapore in July 2026" once the trip ends.
Two numbers are worth knowing. In that announcement OpenAI says it cut the compute needed to serve dreaming to Free users by approximately five times, which is what made the wider rollout practical. The release notes for the same day supply the figure the blog leaves vague: Plus and Pro accounts got twice as much memory capacity. The generations have names, and you will see them quoted elsewhere: saved memories in 2024, saved memories plus Dreaming V0 in 2025, and Dreaming V3 in 2026.
The rollout started with Plus and Pro users in the US, with additional countries and the Free and Go plans promised over the following weeks and an in-product notice when your account is switched over. OpenAI has not published a note saying the rollout is complete, and as of November 2026 its Memory help page says features and controls can vary by plan, region, platform and workspace settings. If your screen does not match this guide, that is the most likely reason.
How to Delete ChatGPT Memory
The current path takes four steps on the web:
- Open Settings, then Personalization, then Memory.
- Open the memory summary page.
- Open the three-dot menu on that page.
- Select Delete and turn off memory.
That single action clears the memories shown on the summary and switches memory off. OpenAI shipped it on 12 June 2026 for all ChatGPT users on the web, according to the ChatGPT release notes, with mobile following, so if you are on an iPhone or Android handset and cannot find it, use a browser.
To remove something without the nuclear option, edit the summary instead. Highlight the line that is wrong and correct it, or type the correction into the text field. For a detail surfaced under a response, act on it from Sources. Don't mention this again is the lightest touch: it reduces future references to the information, but it does not delete the original source.
If your settings look different
If you still see a scrolling list of individual memories with delete buttons, your account is on the legacy saved memories system. That is a supported state, not a bug, and you can move between the two deliberately. Where the option is available, open Settings, Personalization, Memory and select Saved memories to use the legacy list. To return to the new system, select Try improved memory. Switching does not delete your chat history.
The legacy system behaves differently in one way that still trips people up. Saved memories are stored separately from your chat history, so deleting a conversation does not automatically remove a memory created from it, and that memory keeps influencing future answers until you delete it directly. Deleting a saved memory stops it being used, but it does not remove mentions of that information from past conversations. It is also the system that could reach a "memory full" state, which is why so many articles are written around freeing up space.
What deleting does not remove
This is the part that matters most. Clearing your memories does not delete your past chats. If you later switch memory back on, ChatGPT can create new memories from the conversations still sitting in your history, including old ones. People clear memory, feel finished, re-enable it weeks later, and watch the same details reappear.
To remove a detail completely, OpenAI's guidance is to delete every source it appears in: the memory summary, saved memories, regular and archived chats, files saved in your ChatGPT library (or in Space, which replaced Library on eligible Pro, Business and Enterprise accounts from 29 September 2026), and any connected app that holds the same information. Disconnecting an app stops future access, but it does not delete past conversations that already used its content. Deleting a memory also does not rewrite history. If ChatGPT mentioned your city in an old answer, that sentence stays in that conversation until you delete the conversation.
Retention has its own clock. A deleted chat leaves your account immediately and is scheduled for permanent deletion within 30 days, unless OpenAI must keep it longer for security or legal reasons, and archiving a chat does not change its retention period. OpenAI may keep logs of deleted saved memories for up to 30 days for safety and debugging. If your account still shows Reference chat history, turning it off schedules what was remembered from past chats for deletion within 30 days, while the chats themselves stay. Files are separate again: deleting a chat does not delete a file saved separately to Library or Space, and files added to a project stay until the project or the file is deleted. Our guide to what OpenAI collects and who can see it goes through the retention picture in full.
Every ChatGPT Memory Setting Compared
Five different features get called "memory" in conversation, and they behave differently. This is the layout that actually applies:
| Feature | What it holds | Who writes it | Where to manage it | Removed by deleting a chat |
|---|---|---|---|---|
| Memory summary | A synthesis of your chats, files and connected apps | ChatGPT, automatically | Settings, Personalization, Memory | Not necessarily |
| Saved memories (legacy) | Individual noted facts | ChatGPT, or you on request | Settings, Personalization, Memory, Saved memories | No |
| Custom instructions | Rules you write yourself | You, verbatim | Settings, Personalization | No |
| Project-only memory | Context sealed inside one project | ChatGPT, inside that project | Project settings, three-dot menu | Yes, per conversation |
| Temporary Chat | Nothing | Nobody | The chat window, at the moment you open it | Not applicable |
The distinction people get wrong most often is memory against custom instructions. Custom instructions are direct guidance you write and ChatGPT applies as written, which makes them the right home for anything you always want followed. Memory is inferred from conversation and maintained for you. If a preference is non-negotiable, write it as an instruction rather than hoping the summary keeps it.
Memory Inside Projects, Temporary Chats and Locked Chats
Projects, Temporary Chats and locked chats each handle ChatGPT memory differently, and only Temporary Chat keeps a conversation out of memory entirely.
Projects
Projects have their own memory scope. With project-only memory, ChatGPT draws on conversations from inside that project, does not reference your saved memories or conversations outside it, and keeps the project's information out of the memory used elsewhere. Since 14 August 2026 you can switch an existing project between default and project-only memory under Project settings in the three-dot menu, and the change can take a few hours to land. Shared projects use project-only memory, cannot be switched to default, and cannot see a member's own memories. One trade-off: ChatGPT Work is not available in a project set to project-only memory. Our full guide to ChatGPT Projects covers the setup, the file limits and the sharing rules.
Temporary Chat
Temporary Chat is the blunter tool, and since 27 August 2026 you choose how blunt when you open it. Before your first message you pick Personalized or Unpersonalized. An unpersonalized Temporary Chat reads none of your memories, custom instructions or plugins. A personalized one can draw on all three to tailor its answers. Either way it creates or updates no memories, and either way it stays out of your history unless you save it, which converts it into a regular chat that follows your normal account settings from then on. The one thing you cannot do is change your mind: personalisation is fixed once the conversation starts. OpenAI may keep a copy for up to 30 days for safety purposes, so "temporary" describes your account, not OpenAI's systems.
Locked chats
Locked chats hide a conversation, they do not wall it off from memory. OpenAI introduced them on 20 October 2026: a locked chat disappears from your regular history, pinned chats and search, and opens only with a 4-digit account PIN or device authentication such as Face ID. With memory on, a locked chat can still use your regular memory, but new messages sent while it is locked do not personalize your regular chats. Locking does not remove anything ChatGPT remembered before you locked it, and unlocking the chat permanently can feed the whole conversation back into memory. Be careful with the PIN: resetting it permanently deletes all locked chats and any remembered information stored only within them.
ChatGPT Memory and Your Privacy
Sensitive information reaches memory if you share it in a chat, and OpenAI says so directly. Its Memory help page offers no filter that holds back a medical detail or a salary figure. The controls you have are the ones above: switch memory off, open an unpersonalized Temporary Chat, or delete the sources. Since 21 September 2026 the Privacy Center (account menu, Help, Privacy Center on the web) explains each of these controls and links to the setting, though the controls themselves still live in Settings.
Two limits are worth knowing. Turning memory or personalization off does not switch off what OpenAI calls limited safety uses of context in rare, high-risk situations. And on a personal account, OpenAI may use chats and remembered information to improve its models while Improve the model for everyone is on in Data controls.
ChatGPT memory and ads
Memory also feeds advertising. OpenAI's ads documentation is explicit about the link: if ads personalisation is turned on and memory is on, ChatGPT may use your memories and reference recent chats when selecting an ad. The link runs one way only: hiding, reporting, viewing or clicking an ad does not get added to your memory. Advertisers never receive your chats, chat history, memories or personal details, only aggregated figures such as views and clicks. Ads appear on the Free and Go plans and for logged-out users, starting with a US test on 9 February 2026, and on 18 August 2026 OpenAI announced that ChatGPT Ads would expand to 31 European countries the following week. The detail that matters most for memory is the carve-out: personalised ads are not initially available in the EEA or Switzerland, so readers there see ads matched to the conversation in front of them rather than to what ChatGPT remembers about them. We cover every opt-out in how to use ChatGPT without ads. This is the practical reason memory deserves a look even if privacy in the abstract does not interest you: it is the input to personalisation you never explicitly chose.
If you want to see what has accumulated before you decide, prompting is faster than reading settings. Our set of prompts that reveal what ChatGPT knows about you pulls the picture out in a few messages. And if the wider question is how much any assistant retains, stopping AI from training on your data covers the settings across providers.
What Most Guides Get Wrong About ChatGPT Memory
Three claims dominate the search results and all three are out of date.
The first is the deletion path. Many ranking articles still send you to Settings, Data controls, Personalization, Manage memories, then a trashcan icon or a "Clear ChatGPT's memory" button. That is the legacy screen. On a migrated account the route is Settings, Personalization, Memory, and the control is inside a three-dot menu.
The second is the framing of memory as two toggles, Reference saved memories and Reference chat history. Those belong to the older experience. OpenAI's Memory help page now says the controls you see depend on your memory experience, and lists them as possibly including Memory, Reference saved memories, Reference chat history, Memory summary, Manage or Saved memories. A guide that shows only the two toggles is describing one possible screen, not every account's screen.
The third is "memory full". That warning belongs to the legacy saved memories list. In October 2025 OpenAI began managing saved memories automatically for Plus and Pro users, pushing less important ones to the background to avoid the "memory full" state, and its help page says there is no separate storage limit for what ChatGPT can reference through chat history. If you are still hitting that state, you are on the legacy system, and moving to improved memory is a better fix than pruning.
One claim that looks fake but is not: dreaming is OpenAI's own term, not a nickname invented by bloggers. It never appears in the ChatGPT release notes, which is exactly why it reads as invented, but it is the title of OpenAI's own announcement. Treat anything sourced from community forum threads about Dreaming V3 with more caution, since those are user opinions rather than company statements.
Should You Leave It On
For most people, yes. The whole point of the June 2026 rebuild was to stop memory rotting, and a synthesis that revises itself is more useful and less wrong than a list that accumulates. The cost is that you have less granular control than the old system offered, and the summary deliberately does not show you everything.
The sensible middle ground is to leave memory on, read your summary once a quarter, correct anything wrong by highlighting it, and reach for Temporary Chat whenever a conversation is one you would not want shaping later answers. If the context belongs to one piece of work rather than to you, put it in a project with project-only memory instead.
One last thing to plan for: ChatGPT memory has no export button. OpenAI's Memory help page describes no way to download what ChatGPT remembers, so moving to another assistant takes a workaround. Anthropic's import flow for Claude's memory gives you a prompt to paste into your old assistant, asking it to list every memory and every piece of context it holds, and then turns the pasted answer into Claude memory entries. Claude organises what it remembers into topics you can read, edit or delete one by one, which is worth knowing if you are weighing a move; our guide to switching from ChatGPT to Claude covers the rest. If you only want a second opinion rather than a move, Fello AI lets you ask ChatGPT, Claude and Gemini the same question from one app, without a second subscription. Either way, keeping your important context in custom instructions, where you can copy it word for word, is safer than trusting a summary ChatGPT rewrites for you. Read OpenAI's own Memory help page for the reference version of the controls above.
Frequently Asked Questions
How do I delete ChatGPT memory?
Open Settings, then Personalization, then Memory. On the memory summary page, open the three-dot menu and select Delete and turn off memory. That clears the memories shown and switches memory off. It does not delete your past chats, and turning memory back on lets ChatGPT build new memories from the conversations still in your history.
Why does my ChatGPT show a list of memories instead of a summary?
Your account is on the legacy saved memories system, either because the update has not reached it or because it was switched back deliberately. Where the option is available, open Settings, Personalization, Memory and select Try improved memory to move to the new system, or Saved memories to return to the old one. Switching does not delete your chat history.
Does deleting a chat delete what ChatGPT remembers from it?
Not necessarily. Saved memories are stored separately from your chat history, so deleting a conversation does not automatically delete a memory created from it. To remove a detail completely, delete every source it appears in: past chats, archived chats, files, the memory summary, and any connected app holding the same information.
Is ChatGPT memory the same as custom instructions?
No. Custom instructions are guidance you write yourself and ChatGPT applies them as written. Memory is inferred from your conversations and updated automatically. Put anything you always want followed into custom instructions rather than relying on memory to hold it.
Does Temporary Chat use memory?
Only if you choose Personalized. Since 27 August 2026 you pick Personalized or Unpersonalized before the first message, and you cannot change it afterwards. An unpersonalized Temporary Chat ignores your memories, custom instructions and plugins alike; a personalized one can use all three. No Temporary Chat creates or updates memories in either mode. It stays out of your chat history unless you save it, and OpenAI may keep a copy for up to 30 days for safety purposes.
Can I export my ChatGPT memory to another assistant?
Not with a button. OpenAI's Memory help page describes no memory export option. The workaround is a prompt: ask ChatGPT to list every memory and piece of context it holds about you in one block, then paste the answer into the new assistant. Claude has a built-in import flow that supplies that prompt and turns the pasted text into memory entries.
A imagem de destaque deste artigo foi criada com a ajuda de IA.