August has been a month of new places to work. This week Cowork got something that travels between them. Starting August 25, the memory Claude uses in chat is the same memory Cowork uses, so a task you hand to Cowork already knows what you spent months explaining somewhere else. Every topic Claude has saved is now a short file you can read, edit, or delete. And on the developer side, computer use, the Skills API, and the Files API all reached general availability, with a new browser use tool alongside them.
Cowork and Chat Now Share One Memory
On August 25, Anthropic announced that the memory you use in chat is now the same one Cowork uses. When Cowork runs a task in the cloud, what Claude remembers from your chats is there, and what comes up in Cowork carries back to chat. The practical effect is less re-explaining: the context you built up over months of conversations, like your priorities for the quarter or the state of your projects, is available the moment you hand Cowork a task.
Anthropic's own examples are the useful kind. Ask Cowork to draft an update for your manager and it already knows who that is and how she likes updates written. Brainstorm a conference agenda in chat, and when Cowork builds the budget and logistics doc it knows the headcount, the city, and the speakers. Explain once how your team defines its metrics, and every quarterly business review deck Cowork builds after that uses them, with no rebriefing.
Memory also works differently now. Claude adds topics as you chat instead of summarizing a conversation after it ends, so mentioning that a project deadline moved to September means your next conversation already knows it, without you having to say "remember this." You can pause memory or reset it at any time.
Every Saved Topic Is a File You Can Edit
Everything Claude remembers now lives as a list of short files under Topics in Memory settings, where you can read, edit, or delete each one. Because the files are short and now shared across surfaces, a single correction propagates: fix your company's old name in one file and every conversation from then on gets it right.
By default Claude does not save topics it treats as personal or sensitive, including health, race, ethnicity, religious beliefs, politics, and gender identity. If those happen to be the things you are tired of re-explaining, an "include sensitive topics in memory" setting turns them on. With it on, you get a notice each time Claude saves something in that category, saving applies going forward rather than retroactively, and you can switch it off again at any time.
A few categories stay out regardless of that setting: sensitive identification numbers such as SSNs and government ID numbers, criminal history, immigration status, and anything that violates Anthropic's Acceptable Use Policy. Claude tells you when it cannot save something. On availability, memory is on by default on Free, Pro, and Max plans across web, desktop, and mobile, with sensitive topics off by default. On Team and Enterprise, admins control availability and memory stays off for individual users until they turn it on.
Computer Use, Skills, and Files Reach GA
On August 20, computer use, the Skills API, and the Files API became generally available on the Claude Platform, and computer use gained a new browser use tool. This is developer-facing rather than a Cowork feature, but it is worth knowing about because it is the same machinery underneath the things Cowork does for you.
Computer use lets an agent operate software it can see, clicking and typing and scrolling through applications that were never built for automation. The updated tool takes several actions per turn instead of one per model call, so tasks finish in fewer calls and less time, and it is now eligible for HIPAA-regulated workloads under Anthropic's BAA. The new browser use tool extends the same idea to the web: alongside the screenshot it reads the structure of the page, so an agent can act on a specific field or button rather than a position on screen.
The Skills API is the piece that will feel most familiar if you already write skills for Cowork. A skill is a folder of instructions, scripts, and templates that Claude loads only when a task calls for it; the API lets you upload and version your own and attach them to any request, running inside Claude's code execution sandbox with nothing for you to host. The Files API is storage for the documents an agent reads and writes, and it arrives at GA with automatic file expiration, five times higher rate limits, and 1 TB of storage per organization.
Why This Matters for Cowork Users
July gave Cowork more surfaces. August gave it continuity across them. What changed this week is not that Claude remembers - it did that in chat already - but that the remembering is no longer trapped in one app. That makes setup work worth doing deliberately and only once: corrections, preferences, and the way your team defines things become inputs you write down in one place instead of context you re-type at the top of every session. It also makes the Topics list worth an occasional read, because anything wrong in there is now wrong everywhere.
Tip of the Week: Fix the Memory File, Not the Prompt
With one memory across chat and Cowork, the highest-leverage edit is often not a better prompt. It is a correction in Memory settings, because that one edit applies to every session on every surface from then on.
❌ Weak
Re-typing the same preamble at the top of every task: "Remember, we renamed the company last quarter, and our definition of active user is 7-day, not 30-day."✅ Strong
Open Settings, then Memory, then Topics. Correct the company name and the metric definition in the relevant files.
Then just ask: "Draft the Q3 review deck."The rule of thumb: if you catch yourself typing the same correction twice, it belongs in a memory file, not in a prompt.
Worth Reading This Week
- → Claude's memory works everywhere, and you decide what's in it - Claude Blog - The August 25 announcement: one memory across Cowork and chat, the editable Topics list, and the sensitive-topics setting.
- → Use Claude's chat search and memory to build on previous context - Anthropic Support - The help center article on how memory works, including what Claude will not store even with sensitive topics turned on.
- → Build production agents with computer use, the Skills API, and the Files API - Claude Blog - The August 20 general availability post, including the new browser use tool and the raised Files API limits.
- → Claude release notes - Anthropic Support - The running log of what shipped, worth checking before you assume a feature is or is not available on your plan.
Memory is the quietest thing Anthropic shipped this month and probably the one that changes how a Cowork session actually feels. It is worth ten minutes in the Topics list before you hand over your next task.
Until next time.
— The ClaudeCowork.com team