Anthropic’s Claude already had the ability to remember past conversations, but on August 25, 2026, Anthropic introduced a major change to how that memory works. Claude Chat and Claude Cowork now share the same memory system, meaning you no longer have to re-explain project context, preferences, or workflows when moving between the two platforms.
This isn’t a minor update. It solves one of the most frustrating aspects of using AI agents: repeating yourself. If you spend an hour discussing a project, your company’s preferences, or specific terminology in Claude Chat, you can now hand a task to Cowork and it picks up exactly where you left off, with all that context already available.
At a Glance
What changed? Claude Chat and Claude Cowork now share the same memory system.
When? August 25, 2026.
What can users control? View, edit, delete, pause, and reset memory topics.
Sensitive information? Not stored by default. Users can opt in through settings.
Which plans? Free, Pro, and Max plans have memory enabled by default. Team and Enterprise availability is controlled by administrators.
Does Claude Code share this memory? No. Claude Code’s memory system remains separate.
Claude Memory Explained: What Is It?
Claude Memory is Anthropic’s persistent memory system that lets Claude retain useful information from your interactions and use that context in future conversations and tasks. Unlike chat history, which is a record of everything you’ve said, memory consists of relevant information Claude identifies and stores to personalize future interactions.
Think of it this way: chat history is a transcript. Memory is the notes Claude takes while reading that transcript. Claude organizes remembered information into topics, which you can review, edit, or delete anytime through Settings > Memory.
The goal is straightforward: reduce the need to explain the same things repeatedly. If you’ve told Claude about your company’s reporting style, your project deadlines, or your preferred workflow, Claude remembers it and applies that context automatically.
What’s New in the Claude Memory Update
On August 25, Anthropic announced that it is merging the memory systems used by Claude chat and Claude Cowork, meaning information Claude remembers from one experience can now be available when you use the other.
Here’s what changed:
| Update | What It Means |
| Shared Chat + Cowork memory | The memory you use in chat is the same as in Claude Cowork |
| Memory updates during chats | Memories are now created on the fly during chats instead of only after conversations end |
| Topic-based memories | Saved information organized by topic that you can view and manage |
| Edit and delete controls | Users can see everything it remembers “topic by topic,” and edit or delete any of it |
| Sensitive-topic control | Users can choose whether sensitive information is remembered |
| Pause and reset options | Users can pause memory or reset it at any time |
How Does Claude Memory Work?
Claude Memory operates in three stages:
1. Claude Identifies Useful Information
As you chat, Claude identifies information that might help future interactions. This could be project details, your preferences, terminology, priorities, or information that helps personalize responses.
2. Claude Updates Memory Topics During Conversations
Claude now adds topics to memory as you chat, instead of summarizing conversations after they end. This means if you mention your project deadline has moved to September, the next conversation already knows without you having to repeat it.
3. Relevant Memory Is Used Later
When you return to Claude or use Cowork, the system pulls relevant memories and applies that context. You pick up where you left off rather than starting from scratch.
Practical example: You explain your company’s unique approach to defining quarterly metrics in a Chat conversation. Later, every quarterly review deck that Cowork builds uses those same definitions automatically, without you having to re-explain them.
Claude Memory Across Chats and Cowork
This is the core value of the August 2026 update. “When Cowork runs a task in the cloud, what Claude remembers from your chats is there, and vice versa,” says Anthropic. The flow works in both directions.
Chat to Cowork workflow:
Discuss a conference you’re organizing in Claude Chat, mentioning the headcount, location, and speakers. Hand Cowork the task of building a logistics budget and timeline. It already knows all those details and can work from there.
Cowork to Chat workflow:
Cowork generates a quarterly report using your company’s terminology and metrics. Later, when you’re drafting in Chat, that context is available. Claude remembers the framework Cowork used.
Research to execution:
Start research in Chat, building context about a topic over multiple conversations. When it’s time to create something based on that research, Cowork already has the full context. No rebrief needed.
Anthropic told us there’s no option to separate Claude Chat and Cowork’s memories since they now use a shared system. If separation between the two experiences matters to your workflow, the only option is to use different accounts.
What Does Claude Remember?
Claude remembers information likely to be useful in future interactions, including:
- Project context and ongoing work
- Your stated preferences for how work should be produced
- Recurring workflows and processes
- Company terminology and definitions
- Project priorities and deadlines
- Information that helps personalize responses
- Communication preferences
The key distinction: Claude Memory is selective persistent context, not a permanent transcript of everything you’ve said.
Claude doesn’t retain every detail from every conversation. It identifies patterns and useful information while discarding routine back-and-forth or irrelevant details. This selective approach is intentional-it reduces noise while preserving what matters.
What Doesn’t Claude Remember?
Claude has clear boundaries on what it stores:
Claude does not store sensitive information such as health data, race, ethnicity, religious beliefs, political views or gender identity by default.
Additionally, Claude will not save certain information, including government-issued identification, Social Security numbers, criminal history and immigration status. These restrictions apply even if you enable the sensitive-topic setting.
Any information that violates Anthropic’s usage policies is not stored under any circumstances.
Claude Memory Privacy: How Sensitive Information Works
By default, Claude doesn’t store sensitive topics. Users can enable the “include sensitive topics in memory” option if they want Claude to remember such information, and the app will notify them whenever it saves a sensitive topic.
Sensitive topics include health conditions, race and ethnicity, religious beliefs, political views, and gender identity.
Should You Enable Sensitive-Topic Memory?
Enable it if:
You want highly personalized assistance and don’t mind Claude retaining sensitive personal context across conversations.
Keep it off if:
You prefer greater privacy or want to limit persistent context around sensitive topics.
Important: Previously discussed sensitive information is not retroactively added when the setting is enabled. Turning on the setting only affects future conversations.
How to View, Edit, or Delete Claude Memory
Managing Claude Memory takes just a few steps:
How to See What Claude Remembers
Open Claude, go to Settings, then select Memory. You’ll see all your saved memories organized by topic. Each topic is labeled and shows the remembered information.
How to Edit Claude Memory
Find the topic you want to modify, open it, and you can edit the information directly. This is useful for correcting outdated details or adjusting how Claude interprets your preferences.
How to Delete Claude Memory
Select any memory topic and delete it. This removes that information from Claude’s memory immediately.
How to Pause or Reset Memory
In Memory settings, you can pause memory entirely, which stops Claude from creating or updating memories. You can also reset memory to clear your current memory state and start fresh.
Is Claude Memory Available on Free, Pro, and Max?
Yes. Memory is on by default on Free, Pro and Max plans across web, desktop, and mobile.
| Claude Plan | Memory Availability |
| Free | Yes, enabled by default |
| Pro | Yes, enabled by default |
| Max | Yes, enabled by default |
| Team | Admin controlled |
| Enterprise | Admin controlled |
For Team and Enterprise organizations, administrators control whether memory is available to individual users. It’s off by default for these plans unless explicitly enabled.
Important note: Claude Cowork itself is not available on the Free plan, even though memory is. Free users have access to Claude Memory in Chat but cannot access Cowork. Memory availability and Cowork availability are separate features.
Claude Memory Use Cases
For Beginners
Remember basic preferences, reduce repetitive prompts, maintain continuity between chats.
For Professionals
Track recurring project context, remember writing and communication preferences, retain meeting and research context, automate recurring reporting workflows.
For Marketers and Writers
Remember tone and voice preferences, store brand terminology, track recurring content requirements, maintain campaign context across conversations.
For Researchers
Retain ongoing research context and terminology, track project priorities, manage recurring research workflows.
For Businesses Using Cowork
Store organizational terminology and reporting requirements, maintain project context for agent-based tasks, ensure consistent output across automated workflows.
Claude Memory vs ChatGPT Memory vs Gemini Memory
How does Claude’s new shared memory compare?
| Feature | Claude | ChatGPT | Gemini |
| Cross-conversation memory | Yes, shared Chat/Cowork | Yes | Limited |
| User controls | Full view, edit, delete | Limited controls | Limited controls |
| Sensitive topic protection | Off by default, opt-in | Varies | Varies |
| Multi-platform integration | Chat + Cowork unified | Limited agent integration | Limited agent integration |
| Memory updates live | Yes, during conversations | Summarized after | Summarized after |
| Mobile support | Yes | Yes | Yes |
Claude’s advantage is the unified memory system that works across Chat and the agent tool (Cowork), combined with explicit user controls over what’s remembered. You can see exactly what Claude has stored and change it anytime.
Limitations of Claude Memory
Claude Memory is powerful but isn’t perfect. Understand these boundaries:
- Claude may not capture every detail from your conversations. Memory is selective, not comprehensive.
- Stored information can become outdated. You should periodically review and update your memories.
- Incorrect memories can influence future responses. If Claude misremembered something, it will likely repeat that mistake until you correct it.
- Users cannot separate Chat and Cowork memories. The only way to keep Claude Chat and Cowork memories separate is to use different accounts.
- Memory should not replace explicit instructions for high-stakes or critical tasks. Important instructions should be provided in each conversation, not relied upon solely from memory.
- Sensitive-topic memory requires deliberate consideration. Think carefully before enabling it.
Best Practices for Using Claude Memory
Best Practice 1: Review Your Memory Periodically
Set aside time monthly to check what Claude has stored. Look for outdated information or misunderstandings.
Best Practice 2: Correct Outdated Information
If a deadline changed, a project ended, or a preference shifted, update your memories. One corrected memory topic can improve many future conversations.
Best Practice 3: Avoid Relying on Memory for Critical Instructions
For tasks where accuracy is essential, provide explicit instructions in the conversation. Memory should enhance, not replace, clear communication.
Best Practice 4: Keep Sensitive-Topic Memory Disabled Unless Useful
Only enable sensitive-topic memory if you genuinely want Claude to retain that context. Review your privacy preferences regularly.
Best Practice 5: Use Explicit Prompts for Temporary Context
If you need Claude to approach something differently just for one conversation, state it explicitly. Don’t assume memory alone will handle temporary adjustments.
Best Practice 6: Remember That Cowork Shares Memory With Chat
Understand that information you discuss in Chat is available to Cowork, and vice versa. This is a feature for most users but something to keep in mind if you want complete separation.
Should You Use Claude Memory?
Claude Memory is worth using if you:
- Use Claude repeatedly for ongoing work
- Work on long-running projects where context matters
- Follow the same instructions or processes repeatedly
- Use Claude Cowork for agent-based tasks
- Want more personalized responses from Claude
Consider limiting Claude Memory if you:
- Prefer stateless AI interactions with no persistent context
- Frequently discuss sensitive information and want to minimize storage
- Want complete separation between conversations
- Rarely use Claude for repeated tasks
Bottom line: If you use Claude repeatedly for ongoing work, Claude Memory is worth enabling. If privacy and conversation separation matter more, review your Memory settings carefully and enable only what you need.
Frequently Asked Questions
How does Claude Memory work?
Claude identifies useful information during conversations, stores it organized by topic, and retrieves relevant context when you start new conversations or use Cowork. Memory updates happen during conversations, not just after they end.
What does Claude remember?
Claude remembers project context, preferences for how work should be produced, recurring workflows, terminology, priorities, and information that personalizes responses. Memory is selective, not comprehensive.
Does Claude remember previous chats?
Yes, across previous chats. Claude retains relevant information and applies it to new conversations. However, it doesn’t remember every detail-only information Claude identifies as useful.
Does Claude Cowork share memory with Claude Chat?
Yes. The memory you use in chat is the same as in Claude Cowork. Context from Chat is available to Cowork, and information generated in Cowork carries back to Chat.
Can I see what Claude remembers?
Yes. Go to Settings > Memory to view all your saved memories organized by topic. You can read, edit, and delete individual memory topics.
Can I edit or delete Claude Memory?
Yes. You can edit information within any memory topic to correct or update it. You can delete any memory topic entirely.
Can I turn off Claude Memory?
Yes. You can pause memory in Settings, which stops Claude from creating or updating memories. You can also reset memory to clear everything.
Is Claude Memory available on the Free plan?
Yes. Memory is on by default on Free, Pro and Max plans across web, desktop, and mobile. However, Claude Cowork itself is not available on Free.
Does Claude remember sensitive information?
Not by default. Claude does not store sensitive information such as health data, race, ethnicity, religious beliefs, political views or gender identity by default, but users can enable the “include sensitive topics in memory” option if they want Claude to remember such information. Certain information, like government IDs and SSNs, is never stored.
How do I update my app to get the latest Claude Memory feature?
On mobile, update to the latest version of the Claude Android or iOS app to see the new updates.