Key takeaways

  • Claude’s new Memory can remember details from past chats to reduce retyping and context loss.
  • It’s rolling out to all paid users: available now on Max, and coming to Pro over the next few days.
  • Memory was already live for Team and Enterprise since September.
  • You can see, edit, or delete individual memories and even create separate memory spaces for different projects.
  • Import/export is supported, so you can bring memories from other chatbots or take your data with you.
  • Controls live in Settings, with toggles for Search & reference chats and Generate memory from chat history.

What is Claude Memory?

Claude Memory lets the assistant recall helpful details from earlier chats. It keeps track of ongoing work, preferences, and project notes so you don’t have to restate them each time. You stay in control: you can review what’s saved, tell Claude to forget something, or pause memory entirely.

Who gets it and when?

  • Max subscribers: enable Memory today in Settings.
  • Pro subscribers: rollout over the coming days.
  • Team & Enterprise: Memory has been available since September.

Before you start (privacy & control)

  • Transparency: Claude shows exactly what it remembers, not vague summaries.
  • On/off anytime: toggle Memory off for a session or permanently.
  • Edit history: remove items or ask Claude to forget a topic or job entirely.
  • Incognito option: work without saving new memories when you need a clean slate.
  • Separate spaces: use distinct memory spaces to prevent cross-talk between personal and work projects.

The setup (step by step)

  1. Open Claude → Profile → Settings.
  2. Under Preferences, switch on Search & reference chats.
  3. Toggle Generate memory from chat history to let Claude build memory from prior conversations.
  4. (Optional) Create memory spaces for each project to keep contexts separate.
  5. Review memories anytime: open Settings, inspect entries, edit or delete as needed.
  6. To pause Memory for a chat, switch to Incognito before you start.

Smart ways to use Memory (recipes & prompts)

  • Weekly status: “Summarize last week’s tasks and blockers based on our recent chats. Draft a 5-bullet update.”
  • Research roll-ups: “Track sources we cite for my climate report and remind me of gaps each Friday.”
  • Coding continuity: “Remember my local dev setup and test commands. Apply them to new scripts we write.”
  • Client context: “Keep brand voice, glossary, and approvals for ACME Corp in a separate memory space.”
  • Learning plan: “Remember my GRE vocab sets and quiz me with 10 spaced-repetition cards daily.”

Benefits you’ll notice

  • Less repetition: skip the boilerplate; continue where you left off.
  • Fewer mistakes: project details live in memory spaces, not in your head.
  • Faster ramp-up: new sessions inherit context without hunting old threads.
  • Easy migration: import past notes from rival chatbots; export when you’re done.

Best practices (to keep it tidy & safe)

  • Name spaces clearly: “Client-Q4-Ads” beats “work stuff.”
  • Store only essentials: goals, constraints, definitions; avoid sensitive personal data.
  • Prune weekly: delete stale instructions so Claude doesn’t follow outdated rules.
  • Validate first runs: ask, “What do you currently remember about this project?” and correct it.
  • Use Incognito for experiments: keep throwaway prompts out of long-term memory.

Troubleshooting

  • Can’t see Memory? You may be on Pro before the rollout hits your account. Check again later.
  • Irrelevant recall? Open Settings → review entries → delete or rewrite with a short corrective note.
  • Mixing contexts? Move items into the right memory space, then ask Claude to prioritize that space.

Did you know?

You can copy/paste memories from other AI tools to jump-start Claude on day one—and export them later if you switch again. No lock-in.

Conclusion

Claude Memory pushes the assistant from a single chat to an ongoing collaborator. With clear controls, editable entries, and project-based spaces, you get continuity without giving up oversight. Turn it on, seed a few key facts, and let Claude carry the routine context so you can focus on the work.

FAQs

Is Memory available on the free plan?

No. It’s rolling out to paid tiers—Max now, Pro shortly, and already live for Team/Enterprise.

Where do I turn it on?

Profile → Settings → Preferences. Enable Search & reference chats and Generate memory from chat history.

Can I prevent Claude from remembering a specific session?

Yes. Use Incognito to keep a chat out of Memory.

What if Claude stores something I don’t want?

Open Settings to view and delete individual memories. You can also tell Claude to forget a topic in chat.

How do memory spaces help?

They separate contexts—for example, one space per client—so details don’t bleed across projects.

References