ChatGPT memory explained: saved memories, chat history, temporary chats and how to turn it off

ChatGPT memory is how ChatGPT carries what it has learned about you from one conversation to the next: your job, your dietary preferences, the project you are working on, the way you like answers formatted. It saves you from repeating yourself, but it also means ChatGPT is holding personal context you may want to check. This guide explains how memory works as of October 2026, the difference between saved memories and chat history, what temporary chats do, what each plan gets, and how to correct, delete or turn it off. It is based on OpenAI’s Help Center and pricing page. OpenAI says memory controls vary by plan, region, platform and workspace, so your settings may not match ours exactly.

What ChatGPT memory is

OpenAI says that when Memory is on, ChatGPT “can remember relevant preferences and details from your chats and other available sources.” Depending on your plan and region, those sources can include:

• Past chats
• Saved memories
• Custom instructions
• Files in Library
• Content from connected apps, such as Gmail

Memory is not a transcript of everything you have said. OpenAI says it “does not retain every detail from every conversation”, that ChatGPT decides which information is relevant, and that it doesn’t search your history for every response, only when context is likely to help. Memory can also shape web searches: if ChatGPT knows your location or that you are vegetarian, it may use that to refine a restaurant search.

The parts of ChatGPT memory

Improved memory vs legacy saved memories. OpenAI now describes two experiences. Improved memory “continually updates a broader summary of relevant context”, while legacy saved memories use “a more explicit list of individual memory items.” Where the option exists, you can switch in Settings, Personalization, Memory: select Saved memories for the legacy list or Try improved memory to go back. Switching doesn’t delete your chat history.

Saved memories vs chat history. A saved memory is an explicit item, such as “Prefers metric units.” Chat history is looser: ChatGPT draws on past conversations, and OpenAI says what it derives “can change as ChatGPT updates what is most useful to remember.” If your settings show separate switches, turning off Reference saved memories also turns off Reference chat history, but turning off chat history leaves your saved memories in place.

What each plan gets

OpenAI’s pricing page lists memory on every individual plan, with more capacity as you pay more:

OpenAI doesn’t publish how many memories each tier holds. In Business, Enterprise and Edu workspaces, owners and admins decide which memory settings and roles are available, so ask your admin if a setting is missing. For regulated workspaces such as ChatGPT for Healthcare, improved memory is off by default and isn’t covered under a BAA. For help choosing a personal plan, see ChatGPT Go vs Plus and ChatGPT Pro explained.

How to see what ChatGPT remembers

1. Open Settings.
2. Select Personalization.
3. Select Memory.

Depending on your account, you will see some of these controls: Memory, Reference saved memories, Reference chat history, Memory summary, Manage or Saved memories. You can also simply ask ChatGPT, “What do you remember about me?”

When the memory summary is available, you can type a correction in the text field, highlight text and correct it, choose Don’t mention this again, or pick Delete and turn off memory from the more options menu. OpenAI notes that the summary is a high-level view, so it may leave out details that ChatGPT can still use.

When ChatGPT uses personal context in an answer, Sources may appear below the response, pointing to a past chat, saved memory, custom instruction, file or connected email. From there you can open the source, correct or delete a saved memory, delete a referenced chat or mark the source as not relevant. Sources aren’t included when you share a chat by link.

How to correct or delete a memory

• Correct it: tell ChatGPT the information is wrong or out of date, or edit the summary or saved memory directly.
• Stop it coming up: select Don’t mention this again, or ask ChatGPT not to use it. This changes behaviour but deletes nothing.
• Delete it: remove the item from the memory summary or saved-memory list.

Deleting is where people get caught out. OpenAI says “deleting a chat alone does not necessarily delete a separate saved memory created from that chat”, and deleting a saved memory doesn’t remove mentions of it from past conversations. To remove something completely, delete it from every place it lives: the memory summary, saved memories, regular and archived chats, files in Library and connected apps. OpenAI says deletion can take time to propagate and that it may keep logs of deleted saved memories for up to 30 days for safety and debugging. Our guide to deleting AI chat history covers clearing chats in ChatGPT, Gemini and Claude.

How to turn ChatGPT memory off

Go to Settings, Personalization, Memory and switch Memory off, or use Delete and turn off memory from the summary’s menu to clear what the summary shows at the same time. Things to know:

• Your chats stay. Turning Memory off doesn’t delete past chats, and if you turn it back on later, ChatGPT may build new memories from chats still in your history.
• Chat history has its own 30-day clock. If you turn off Reference chat history, information remembered from past chats is scheduled for deletion from OpenAI’s systems within 30 days. The chats themselves remain unless you delete them.
• Safety use continues. OpenAI says turning off memory doesn’t stop “limited safety and security uses of context in rare, high-risk situations.”
• Teen accounts: a parent or guardian can manage supported memory controls through Parental controls.

Temporary chats and memory

A temporary chat stays out of your history and, while it remains temporary, isn’t used to train OpenAI’s models. To start one, open a new chat, select Temporary, then choose:

• Personalized: can use your existing memories, custom instructions and plugins.
• Unpersonalized: uses none of them.

You have to choose before your first message and can’t change it afterwards. Either way, a temporary chat never creates or updates memories. OpenAI may keep a copy for up to 30 days for safety purposes. If you save a temporary chat, it becomes a regular chat and from then on follows your normal memory and model-improvement settings.

Memory in projects and other ChatGPT tools

Projects can keep memory to themselves. In a project’s settings, under Memory, choose Default memory or Project-only memory, then Save; OpenAI says changes may take a few hours. With project-only memory, your saved memories aren’t used in the project, and its chats can reference each other but not chats outside it (and vice versa). Shared projects always use project-only memory. On Plus and Pro, ChatGPT can reference previous chats within a project and prioritises the project’s chats and files. There is no list of project memories; to stop ChatGPT using a conversation, delete it or move it out.

Elsewhere, memory doesn’t always follow you. OpenAI says ChatGPT memory and skills don’t carry over to the ChatGPT add-in for Word, whose conversations are kept separate. Pages you share in ChatGPT Space don’t expose your Memory, though anything ChatGPT writes onto a page is visible to its readers. And Pro subscribers’ ChatGPT dots receive memories from ChatGPT and form their own.

Privacy and training

Memory holds whatever you tell ChatGPT, including sensitive details. OpenAI’s advice is to review your Memory settings and Sources, and to use an Unpersonalized temporary chat when you don’t want a conversation to use or create memories. For personal accounts, OpenAI may use chats and remembered information to improve its models when Improve the model for everyone is on; you can switch that off in Data controls. By default, OpenAI doesn’t train on content from Business, Enterprise, Edu or ChatGPT for Healthcare workspaces.

Should you keep memory on?

• Keep it on if you use ChatGPT daily for the same work, studies or household planning, and check the summary every few months.
• Use project-only memory for clients or topics you want kept apart.
• Use temporary chats for one-off questions about health, money or anything you would rather it didn’t keep.
• Turn it off if you share a device or account, or simply prefer every chat to start fresh.

Other assistants handle memory with their own settings; our ChatGPT vs Claude comparison and the ChatGPT profile are good next reads.