Enjoying this issue?
Get tomorrow's AI & engineering digest in your inbox — hand-picked, summarized, and always spam-free.
TLDR
Anthropic has released a unified memory system for Claude that now works across both Chat mode and Claude Code's Co-work mode when running tasks in the cloud, ending the previous limitation where Co-work had poor recall. Users can see, edit, or delete memories topic by topic, and the system adds memories mid-conversation rather than summarizing at the end. The practical trade-off: this only functions when Claude is running in the cloud, not when running locally on a user's machine.
Key points
Anthropic released a new Claude memory system that syncs memory between Claude Chat and Claude Co-work (Cloud Code) for the first time, solving the previous issue of Co-work having very weak memory.
Memory is now persistent and topic-based; Claude adds facts to memory during a conversation as they are discussed, rather than waiting until the end to summarize.
Users can view, edit, or delete specific memory entries by topic (e.g., preferences, profile, specific projects). Memories are organized as editable files within a memory settings panel.
The new memory system is enabled by default for all users on Free, Pro, and Max plans, across web, desktop, and mobile—no opt-in required.
Claude does not automatically store sensitive topics like health or beliefs to memory, but users can manually turn on inclusion of those topics in memory settings.
There is no shared memory between Claude Co-work and Claude Code (the CLI tool) at launch; the presenter speculates this integration may come later.
A critical caveat: the shared memory only works when Claude Co-work tasks are run in the cloud (in beta), not when run locally on the user's computer. If running locally, the memory sync does not function.
Users can import memories from other AI providers (ChatGPT, Gemini) directly within the memory settings interface.
Tools mentioned
Techniques
- Memory persistence across chat and cloud-based Co-work modes via shared memory system.
- Topic-based memory organization with user-editable records.
- Mid-conversation memory capture (storing facts as they are mentioned rather than summarizing at end).
Stop scrolling. Start reading smarter.
Receive the day's most important AI & engineering updates in one concise email. No spam.
Transcript (captions)
Enthropic just dropped a new Claude memory system that basically combines the memory of Claude Co-work and Claude chat into one. And if you've ever used Claude Co-work, one of the negatives
about it is it had very bad memory. Well, this solves it. So, in this video, I'm going to break down exactly what's new, how to turn this on inside of your account. And then I'm going to break
down one big caveat that we need to talk about so you could fully understand how this works. Let's dive in. All right. So, first things first, let's dissect this article that Enthropic just
released talking about the new shared memory between Claude Co-work and Claude Chat mode. There's a few key things here that is actually really interesting and then not only am I going to talk about
this, but I'm going to dive into Claude Co-work and show you exactly how this works. First of all, the headline for this is Claude's memory works everywhere and you decide what's in it. This is
kind of the main headline right here. the fact that we could decide what Claude actually remembers. And I'm going to show you exactly how to do this later on in this video. So, make sure to stick
to the end. Now, starting today, the memory you use in chat is the same as Claude co-work. Now, wherever you work with Claude, it starts from what it already knows about you. If you've used
Claude Co-work in the past, unless you're working in a particular project, it usually doesn't remember you or everything about your business in the previous conversations you had. And this
is one of the reasons why a lot of people were only using Claude chat mode instead of co-work because we didn't have this persistent memory on us in our previous conversations. Now it says you
can see everything Claude remembers topic by topic and edit or delete any of it. So it almost acts as a second brain now directly inside of Claude that we have access to. And again, we can make
changes to what it remembers. So we could kind of enrich these memories that it has on us. And this is an interesting point right here for privacy concerns. Claude does not store subjects
considered sensitive like health or beliefs to memory by default. But if those are topics you are tired of re reexplaining, you can turn them on inside of the memory settings. So it's
cool that we could customize you know the thing the personal things that Claude remembers about us if we want to give it more context which for me I have no problem doing. One memory across
claude co-work and chat. Co-work now has memory and it's the same one that you use in chat, leading to less explaining and picking up where you left off. When Co-work runs a task in the cloud, what
Claude remembers from the chat is there and vice versa. The context you've built up across months of conversations, for instance, your Q3 priorities and the status of your projects is there the
moment that you hand co-work a task and what comes up in co-work carries back to chat. So, this works both ways. If we're interacting with Claude Co-work and we say something in particular that maybe
we're working on for the day, when we go back over to Claude chat mode, it's going to have that persistent memory there as well. Now, this is a pretty interesting point here. The way that it
creates these memories is different than in the past. So, let me break this down right now. Claude now adds topics to memory as you chat. instead of summarizing a conversation at the end.
So if we're in the middle of a conversation, we say something specific that Claude needs to remember, it will know that it needs to store that as a memory instead of waiting for the whole
conversation to go on and then coming up with the TLDDR points that they that we need to save as a memory. So this seems like it'll be much more efficient and we don't have to say things like remember
this. It's just automatically going to add these different memories to our system. Now this is the point that's probably the most interesting about this whole release. We can see and edit our
saved memories as you can see here. So in this example, it shows a couple of different key topics here. So it shows your profile, your preferences, topics, specific things like meal planning,
client pitching, half marathon training, etc. And we could even click on these files and go and edit them manually to change certain things about them. Or we could delete them if it's sensitive
information that we don't want Claude to have on us. And then this is what it looks like when we click on those memories. So, we can just chat with it in this simple chat user interface and
tell Claude exactly what it is that we would like to change. Or I could imagine we could literally click inside of here and make changes to this like it's a markdown file. Now, this is great news
for everybody because this is now on by default for the free plan, pro plan, and the max plan across the web, desktop, and mobile. So, basically, everybody's going to have access to this no matter
what plan you're on. Now, I did see this reply on Twitter, so I figured I would address this point right here. This guy is asking, "What about claude co-work and claude code and vice versa?" As of
right now, this is only syncing between co-work and chat mode. There is no way for us to have a shared memory between co-work and claude code. But if I had to guess, this is probably something that
they're working on that will probably be rolling out over the coming weeks or months. But that's just my speculation. All right, so now let me show you exactly where we can find this inside of
our Claude account. First of all, if I come to cloud.ai AI in my browser. You can now see when I open my account, I'm going to see this right here. So, it shows that Claude remembers co-work
tasks too. We can see all the memories that it's saved as a file. And then it shows that we have more control over specific sensitive topics. So, I could either click not now if I don't want to
enable this, but I'm just going to go ahead and click save preferences. For me personally, I'm just going to toggle this on so it includes sensitive topics inside of memory, but obviously do your
own due diligence and maybe toggle that off if you're skeptical. So, I'm just going to quickly now select save preferences. And this is the interface that we're going to see inside of
Claude. And obviously, we have chat mode right here, and co-work, which now have a which now have a shared memory across both of these products. Now, we're going to see the exact same thing inside of
the Claw desktop app, which you're looking at right here. So, again, I'm going to have this exact same prompt here that I need to fill out, and I'm just going to go and save my
preferences. And now, let me show you how to find our memories that Claude already has stored. So, if I come down to my profile at the very bottom, I'm going to click on this account. I'm just
going to click on this and then click on settings. And if I scroll down to the very bottom, we're now going to see this memory button right underneath our plugins. So, I'm going to click on
memory. And here are some of the memories that it saved for me over the past couple of months. And as of right now, these are just my chat memories. We're going to see a few things. So,
first of all, I could come here and click import memory from other AI providers. So, if I want to import our memory from ChachevT or Gemini, for example, I can do that right here. Next
up, it breaks these down by categories, which I find is pretty cool. So, for example, it shows my preferences, which is one specific file here. Shows my profile. And then here are all the
different topics. So, it has my clothing fit preferences. It has my school community topics. It has finances, fitness, photography, podcast, reading, tools. All of these are broken down by
topics. And as I interact with Claude more and more on both co-work and chat mode, these topics are going to expand and the people that I work with or mention quite often. As you can see,
Aiden right here is my business partner and it remembered him. Now, let me click on one of these to show you what it looks like. So, if I click on clothing, this is a very simple memory. It just
shows specific hoodie fit that I like and it even shows the details of me and the types of men's hoodies that I like. So, for example, this is pretty funny. It shows that I only want men's hoodies
that aren't boxy. And it even remembers me giving a complaint saying, "I bought an expensive hoodie, but the wrist cuffs in the waist cuffs were stretched out, and I want to make sure this doesn't
happen." So, let's say I want to change this memory. I could just come here and say, "I also want to mention that for my shirts, I've really been into a little bit more oversized around the arms, but
I don't like them to be too long. I also like just more cotton fabric." So, obviously, I'm just giving a brain dump. I'm clicking send. And now it's going to filter this information into this
specific memory file. And just like that, at the very bottom, it now added this detail for how I like my shirts. Now, there is something really important that we need to talk about, and that is
that all of these memories are saved everywhere except right here. So, if we turn on So, if we turn on this new memory system, this works inside of Cloud Co-work in chat mode, but only in
the cloud. But if we're running Claude Co-work on our computer locally, then it's not going to have this same shared memory system. Let me explain this further. If you don't fully understand
the difference between it working locally on your computer versus working in the cloud. So, if I come back into Claude Co-work right here at the very top right, you're going to see this
little desktop icon, and I'm going to go ahead and click on this. There are two different ways that we could run this task. Either we could run it in the cloud which is in beta which basically
keeps everything going even when we close the app on our computer or we could close our computer for example. But if we have this running on our computer this only runs when our
computer is on. So in order for us to have this shared memory across cloud co-work and chat mode we need to be running this in the cloud. Otherwise it's not going to work. Now, this isn't
the biggest of deals because as you can see here, I'm running this task in the cloud and I could still go and select the different folders on my computer. So, for example, let me select my
YouTube's video project and folder that I have on my local desktop and I could just come here and say, hey, what is my shirt preference? I know that has nothing to do with YouTube videos, but
I'm going to send that off and it's going to show that that memory that we saved earlier inside of Claude chat mode will then translate and we could access that inside of co-work. And just like
this, it's searching the memory here. It's pulling up the clothing memory. And here is exactly what it found. A bit more oversized around the arms, not too long in length in more cotton in the
fabric, which is exactly the memory that I saved earlier on. Let's test this on the co-work side and give it a memory to create and then see if it translates over into chat mode. All right, so as
for my pants, I kind of like my pants slightly oversized these days. Can you save that to the memory as well? Here we go. It says it added this to the memory. And if I click on the memory, you can
now see that it also added these details for how I like my pants. So, let's just come back to chat mode and say, "Hey, do you know how I like my pants styled?" So, here we go. It just pulled from that
memory file and this is syncing between cloud co-work and chat mode. But again, this has to be running in the cloud. Now, not only will it store this memory inside of the desktop app like you're
seeing right here, but if I pull up my phone, it would also save this memory there as well. Anyways guys, there you have it. If you enjoyed the video, leave a like, leave a comment, and subscribe
for more content like this. This channel is AI for non techies, where I teach non techies how to use AI. Thanks for saying the end, and I'll see you in the next video. Peace.