Connect ChatGPT and Claude
Ask about the people in your Ember right inside ChatGPT or Claude.
Drafting an email in ChatGPT, you ask for Minh's wife's name and keep writing.
Keeper gets more useful as it learns your preferences and the names you use. Memory review puts all of it on one page, so you always know what it knows and can change any line.
Keeper assumed your colleague Minh is your brother because you talk about him often. You open memory review, see the line marked as a guess, and delete it. The next answer gets him right.
Each item says whether you told Keeper, you corrected it, or Keeper guessed. Guesses are easy to spot, so nothing quietly turns into a fact about you.
Tell Keeper to keep replies short or to call your mother "Mum", and it saves that as an instruction. The preference applies to every conversation after.
Earlier chats stay in a history you can search and reopen, so an answer you got last month is still there when the question comes up again.
Keeper can infer things while you chat, and those items are labelled as guesses so you can check them. You can edit or delete any memory at any time.
Deleting a memory makes Keeper forget it permanently. It will not use that item in later answers unless you tell it again.
No. Keeper's memory belongs to your account and is only used to answer you. The people you talk about are never told what Keeper knows.
Ask about the people in your Ember right inside ChatGPT or Claude.
Drafting an email in ChatGPT, you ask for Minh's wife's name and keep writing.
Ask in plain words, and Keeper finds what you saved: an address, a promise.
"Where does David live again?" You ask on the way over, and the answer is already there.
Say one sentence. Keeper writes it down and files it on the right person.
"Lan just got promoted." By the time you open the car door, it is on Lan's page.