Auto memory
Notes writes for itself from your corrections and preferences. See what it is, how it works, how it is used and what to watch for, with pictures and examples.
What it is
Auto lets keep notes for itself across sessions without you writing anything. As it works, Claude saves notes about you, the corrections you give, ongoing project decisions, and where to find outside information. It skips things it can read from the and things your CLAUDE.md already says. Each gets one memory folder, shared by all its , and it stays on your machine.
How it works
-
You correct or guide Claude
-
Claude saves a short topic file
-
MEMORY.md lists every
-
Next session loads the index
How to use it
-
~/.claude/projects/<project>/memory/the folder with the
MEMORY.mdindex and one file per memory. -
/memorybrowse and open memory files, or turn auto memory on and off.
-
Ask "remember that…" — for example "always use pnpm, not ", and Claude saves it to auto memory.
-
autoMemoryEnabledset to
falsein a project's settings to turn it off for that project.
Watch out
Only the first 200 lines or 25KB of MEMORY.md load at session start; other topic files are read only when needed. Memory is , not enforced rules, so use a hook when something must be blocked. Auto memory is what Claude writes, while CLAUDE.md is what you write.
Related terms
Original and official docs
This page explains the term in plain words. The original is the authority: if they differ, follow the original.
- See this term in the originalClaude Code Docs — Glossary Anthropic · checked 2026-10-19
- Anthropic — Claude Code docs: How Claude remembers your project — Auto memory(opens in a new window)
- Anthropic — Claude Code docs: How Claude remembers your project — Claude md vs auto memory(opens in a new window)