If you’ve ever found yourself re-explaining the same project details, preferences, or context to Claude across different devices or sessions, that friction is finally gone. As of August 2026, Anthropic has rolled out a major update that syncs Claude’s long-term memory seamlessly across every workflow—web, mobile, desktop, and now Cowork in the cloud.

Here’s what’s changed: whatever Claude learns about you—your working style, project specifics, deadlines, preferences—is no longer trapped in one conversation or one device. It lives with your account and shows up wherever you pick up the thread.

Say you mention in a mobile chat that a client deadline moved to Friday. The next time you open Claude on your laptop, whether in a fresh chat or handing off a task in Cowork, it already knows. No copy-pasting, no re-typing context, no “wait, let me find that earlier message.”

This isn’t just convenience—it’s a fundamental shift in how AI assistants integrate into real work. Instead of treating every session as a blank slate, Claude now builds a persistent, evolving understanding of you and your work over time.

How it actually works

Claude doesn’t store full transcripts of your conversations. Instead, it synthesizes them—extracting relevant details like your role, communication preferences, project milestones, and recurring topics, then storing those as discrete memory “topics.”

There are two ways memories get created:

Automatic memory: As you chat, Claude picks up on patterns and details you mention repeatedly—like your industry, preferred response length, or tools you use—and adds them to memory over time.

Intentional memory: You can tell Claude directly to remember something. Say “remember that I prefer bullet points for summaries” or “add to memory that I’m working on a Q4 product launch,” and it updates immediately.

These memories are organized by topic, not as one giant summary, which makes them more precise and easier to manage. You can view, edit, or delete individual memories at any time in Settings > Memory on claude.ai or the mobile app.

Synced everywhere you work

Before this update, there were gaps. Memory worked on the web and mobile, but Cowork tasks in the cloud didn’t always carry the same context. Desktop apps on different machines sometimes fell out of sync.

The August 2026 update closed those gaps. Now:

• What you build in a chat carries into Cowork tasks you dispatch in the cloud.
• What happens in a Cowork task feeds back into your main chat memory.
• Your memory is tied to your Claude account, not the device or surface you’re using.

That means you can start a brainstorming session on your phone during a commute, hand off a structured task to Cowork while at your desk, and pick up the thread again on desktop later—all with the same contextual continuity.

You decide what’s in it

One of the most important design choices here is user control. Memory is on by default, but you’re not locked in. Anthropic has built in multiple layers of control:

• View and manage: See everything Claude has stored, organized by topic, with the ability to delete individual items or reset entirely.
• Pause temporarily: Turn memory off for a period without deleting what’s already stored.
• Disable entirely: Turn off “Generate memory from chats” and “Search and reference chats” in Settings if you prefer a fresh-start model every time.
• Incognito chat: Start individual conversations in Incognito mode (via the ghost icon or Ctrl+Shift+I) where no memory is created at all.

There’s also a privacy safeguard: by default, Claude does not store sensitive subjects like health conditions, race, ethnicity, gender identity, religion, or political beliefs unless you explicitly opt in via a toggle.

Why this matters for real work

For people who use Claude as a daily work partner—writers, developers, researchers, product managers—this update removes a lot of invisible friction. You’re no longer the sole keeper of context. Claude becomes a collaborator that remembers what matters, so you can focus on the work itself instead of re-establishing the groundwork every session.

It also signals a broader trend in AI: the move from stateless, one-off interactions toward persistent, personalized augmentation. Other assistants are heading in similar directions, but Anthropic’s approach stands out for its emphasis on user control and transparency about what’s stored and why.

Getting started

If you want to see what Claude already knows about you, just ask: “What do you remember about me?” It’ll list your current memories right in the chat.

To dive deeper, head to Settings > Memory in claude.ai or the mobile app. You’ll see a full list of stored topics, each with options to edit or delete. From there, you can fine-tune what stays, what goes, and how you want Claude to use it going forward.

For teams using Cowork, the synced memory means project-specific context stays focused and relevant within each project’s dedicated memory space, separate from other projects or general chats.