Custom Instructions are explicit guidance you write for ChatGPT about what it should know or how it should respond. Memory uses relevant context from prior interactions and other available sources to personalize replies. Both can shape responses, but they work differently—and switching a control off does not always delete information already stored elsewhere.
How Memory and Custom Instructions differ
OpenAI describes Custom Instructions as direct guidance supplied by the user. Memory is context that can develop across conversations and, depending on the account and settings, may draw on saved memories, past chats, files in Library, Custom Instructions, or connected apps. ChatGPT does not retain every detail or use every available detail in every response; it determines what context is relevant. See OpenAI’s Memory in ChatGPT and Custom Instructions documentation.
| Setting | What it does | How it changes or persists |
|---|---|---|
| Custom Instructions | Lets you state what ChatGPT should consider, including what it should know or how it should respond. | You can edit or disable them. OpenAI says changes apply immediately across chats, although older conversations may still show earlier versions in their history. |
| Saved memories | Stores details you ask ChatGPT to remember, or details it saves as useful context when the feature is available. | They are stored separately from chat history. Deleting the conversation where a memory began does not by itself delete that saved memory; you can review, correct, or delete memories where controls are available. |
| Reference chat history | Allows ChatGPT to use relevant information from past conversations to personalize future responses. | Context derived from chats can change. Turning this control off schedules information remembered from past chats for deletion within 30 days, but does not delete the original chats or saved memories. |
Which setting should you use?
Use Custom Instructions for directions you want to state directly
If you have a response preference—such as wanting concise explanations or a particular format—Custom Instructions are the direct place to express it. They are user-authored guidance, not a guarantee that every instruction will be followed in every response.
Use Memory for useful context that develops over time
Memory can help ChatGPT use relevant details from prior interactions without requiring you to repeat them. It is not a complete transcript or a promise that every discussed detail will be remembered. ChatGPT chooses what context is useful, and the available sources and controls depend on your account.
#1 Best Overall
Where to find the controls
Memory
OpenAI’s Memory documentation points to Settings > Personalization > Memory. Depending on your Memory experience, you may see controls such as Memory, Reference saved memories, Reference chat history, Memory summary, Manage, or Saved memories. OpenAI says availability can vary by plan, region, platform, and workspace, so labels may differ or some controls may not appear.
Custom Instructions
On web and desktop, the documented path is Settings > Personalization. On iOS and Android, it is Settings > Customize ChatGPT. OpenAI’s instructions describe enabling customization and entering guidance, or disabling it; you can also delete text from the fields. The published character limits are 1,500 for Free and Go users and 5,000 for Plus, Pro, Enterprise, Business, and Education users. These are the limits stated in OpenAI’s help article and may change.
What turning a setting off does—and does not do
- Turning off Memory does not delete past chats. If you turn Memory on again, ChatGPT may create new remembered information from conversations that remain in your history.
- Turning off Reference chat history is not the same as deleting conversations. OpenAI says information remembered from past chats is scheduled for deletion within 30 days. The original chats remain unless you delete them separately, and saved memories are not removed by this control.
- Deleting a chat alone may leave a saved memory intact. To remove a detail saved as a memory, OpenAI says to delete that saved memory and the chat where it was first shared. You may also need to remove the information from other relevant sources, such as other chats, Library files, or connected apps.
- Deleted-memory logs may persist temporarily. OpenAI says logs of deleted saved memories may be retained for up to 30 days for safety and debugging.
- Disabling personalization does not eliminate every safety use of context. OpenAI says limited safety and security uses may continue in rare, high-risk situations.
How Temporary Chat interacts with both settings
Temporary Chat is a per-chat choice. Before it begins, you can choose whether existing Memory and Custom Instructions apply. Temporary Chats do not create or update memories, even when personalization is on. If you want to control whether existing personalization affects a particular Temporary Chat, make that choice before starting the conversation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Does ChatGPT check your history for every answer?
No. OpenAI says ChatGPT uses relevant context when it helps personalize a response; it does not search or retain every detail from every conversation for every reply. Memory is selective, and the available context depends on your settings and account experience.
Recommended Free Tools
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
