ChatGPT Temporary Chat Now Offers a Personalization Choice Before You Start
ChatGPT’s temporary chat mode has changed. Starting 27 August 2026, users can choose whether a temporary conversation should remain fully non-personalized or whether it should be allowed to use saved memories, connected plugins, and custom instructions.
Temporary chat has long been the option people used when they wanted a conversation that felt separate from their regular ChatGPT history. It was designed for quick questions, private brainstorming, or one-off tasks where ChatGPT did not rely on what it already knew about the user. That is still the default behavior, but there is now an important new choice at the beginning of the conversation.
When opening a temporary chat, a small menu appears in the top-right area of the empty chat window. By default, it is set to “Non-personalized.” In this mode, ChatGPT does not use memory, custom instructions, or plugins. If you ignore the menu and start typing, the temporary chat works much like before.
The new option is “Personalized.” If selected before the first message, the temporary chat can access your saved memories, your custom instructions, and connected plugins. This means ChatGPT can give answers that better match your preferences, writing style, work habits, or previous instructions, while the chat itself still starts as a temporary conversation.
There is one major catch: the choice must be made before the conversation begins. Once you send the first message, you cannot switch that temporary chat from non-personalized to personalized, or the other way around.
Another useful change is the ability to save a temporary chat. If a conversation starts as temporary but later becomes important, you can move it into your regular chat history. Once saved, it is no longer treated as temporary. From that moment, your normal ChatGPT account settings apply, including settings related to whether your content may be used to improve models.
If memory is enabled, a saved chat may also influence future responses. However, a personalized temporary chat has a clear limit while it remains temporary: it can read existing saved memories, but it does not create new memories during that temporary session.
Privacy-focused users should also note that temporary does not always mean instantly deleted. For safety purposes, OpenAI may keep a copy of temporary chats for up to 30 days, regardless of whether the chat is personalized or non-personalized. Safety systems may also use limited context from earlier conversations in rare high-risk situations to provide a more appropriate response.
The feature has been listed for paid plans such as Plus, Pro, and Business, but some free users have also reported seeing the same personalization menu. If you use ChatGPT regularly, it is worth checking your temporary chat options, memory settings, custom instructions, and active sign-ins to make sure they match your privacy preferences.
In short, ChatGPT temporary chat is no longer just a one-size-fits-all private mode. It now gives users more control: keep the conversation completely separate, or allow ChatGPT to bring in your saved preferences for more personalized answers. The default remains non-personalized, but the new setting makes it important to look before you type.
