Claude memory you can read: keep it in your notes

Claude memory lets Claude carry context from one conversation into the next, so you do not have to explain yourself every time. It is useful. It is also not the only kind of memory worth having. Some things you want Claude to know are better kept somewhere you can open, read and change yourself: in your notes.

How Claude's own memory works

According to Claude's help center, memory lets Claude build up an understanding from your chats and bring it into new ones. It saves what it learns as a set of individual topics while you talk, and you can also simply tell it to remember something.

• Plans. Memory is on by default on the Free, Pro and Max plans, on the web, desktop and mobile. On Team and Enterprise plans it is off until an owner turns it on.
• Control. Under Settings, Memory, you can turn it on or off, open any saved topic, read it and edit it.
• Projects. Each project in Claude has its own separate memory.
• Incognito. Incognito chats are not remembered.
• Chat search. On paid plans, Claude can also search your earlier conversations when you ask it to.

That is a fair amount of control, and for the everyday "remember that I am vegetarian" kind of fact it does the job well. So why bother with notes?

Where memory in notes is better

Claude memory is Claude's understanding of you, stored inside Claude. A note is a thing you wrote, stored where you keep your own stuff. The difference shows up in a few places.

• You use it too. A note with your running plan is something you read at the start of every run. A memory topic is something you visit in settings now and then.
• It holds more than facts. Memory is good at short things about you. A recipe, a packing checklist or a twelve-week plan is a document, and documents live in notes.
• It is exact. A note says precisely what you wrote. There is no summary in between, and no guessing which topic something ended up under.
• It travels. Notes do not belong to one assistant. The same note can be read by Claude today and by another agent next year, and by you on your phone in between.
• It works outside Claude. You can open the note in the app with no chat at all.

None of this makes Claude memory bad. They do different jobs. Let memory hold the small facts about you, and let notes hold the things you and Claude both work on.

What to keep in a note for Claude

• Preferences with detail. "Cooking: vegetarian at home, fish is fine when eating out, no coriander, dinner for two on weekdays."
• Plans in progress. A training plan, a study schedule, a garden plan for the season.
• Lists that change. Shopping, packing, books to read, gift ideas, things to fix in the house.
• Decisions you do not want to make twice. Which paint colour you chose, what the plumber said, why you picked one insurance over another.

Keep one topic per note and give it a plain title. "Cooking preferences" beats "misc". A title Claude can match is half the work.

How Claude reads and updates those notes

Claude can only use notes it can reach. Connect your notes to Claude once, as described in Claude connectors for your notes, and then just ask:

• "Read my note Cooking preferences before you suggest dinners this week."
• "Update my running plan: I missed Tuesday, shift the rest of the week by a day."
• "Add what we just decided about the kitchen tiles to my note House decisions."
• "Check my packing list note and tell me what I have not ticked off."

Naming the note matters. Claude looks for an existing note before creating a new one, but "my note Cooking preferences" leaves no doubt. You can also put a line in Claude's own memory, such as "my cooking preferences are in my notes", so the two kinds of memory point at each other.

Keep an eye on both

Memory in notes has one simple advantage when something is wrong: you open the note and fix it. Do the same with Claude memory now and then. Look through the saved topics, correct what is stale, and move anything long into a note where it belongs.

Claude's help center also describes an experimental way to import and export memory between Claude and other AI services. Notes need no export step. They already sit in a place that any assistant you connect can read, and that you can read yourself without any assistant at all.

Your notes stay yours either way. You can see every connected assistant in the app under Account and revoke it at any time. Reading your notes never needs a subscription. Writing does.

Very Simple Notes is a simple notes app made for this kind of shared memory: plain notes and checklists you write in seconds, that Claude can read and update once you connect it. Hosted in the EU, no ads, and your notes are not used to train AI models. It costs $0.99 a month, and the first month is free.