Prompting

Claude Memory Is Now Files You Can Open And Edit

What to put in Claude's memory on day one, how to correct it when it gets you wrong, and the one setting that decides whether your tasks can see any of it

5min read
13numbered steps
8sections

What actually changed

On August 25, 2026 Anthropic changed how Claude's memory works. It is worth being precise about this, because "the AI remembers you" has been claimed by everyone for two years and this is a different thing from what the others shipped.

Memory used to be a single rolling summary. Claude finished a conversation, wrote a paragraph about you, and that paragraph was the memory. You could not see inside it, you could not correct it, and it was written after the fact.

It is now a set of individual files, grouped by topic, and Claude adds to them while you are still talking rather than summarising once you are done. You can open any one of them, read it, change a line, or delete it.

  • Settings, then Memory, then Topics
  • One file per subject, not one paragraph about you
  • Written during the conversation, not after it
  • On by default on Free, Pro and Max, across web, desktop and mobile
  • Chat and Claude Cowork now share the same memory

Keep reading

Free, for an email.

Unlocks 7 more sections, the document version, and every other guide on the site, for free. Enter your email once to keep reading.

19,000+ follow where these guides come from.No spam. Unsubscribe in one click.

Read next