Transcription
The Claude you're using right now and the Claude I'm using right now are not the same tool. There are five features buried in the settings that most users have never touched, and turning them on makes Claude around 10 times more powerful. So, in this video, I'm going to walk you through every single one of them, and the last setting honestly feels like it's from the future.
Claude has a full memory system, and a lot of users don't know it exists because it's disabled by default. To turn it on, I'll click on my name in the bottom left corner, go to settings, then capabilities, and right there I can see an option called "generate memory from chat history." I'm going to turn that on, and what this does is it lets Claude remember things about me across conversations. So, if I tell Claude that I prefer short paragraphs with no bullet points, or that I always want code examples in Python, it stores that and applies it to every future conversation without me having to repeat myself.
And the memory actually works on three levels. The first level is conversation memory, which is what Claude picks up within a single chat. The second level is project-specific memory, which stays inside a project and carries across every chat within that project. And the third level is global memory, which follows me everywhere across all of Claude.
So, to set a project memory, I need to create a new project and add my instructions in the project description. I'll create a new project, and in the description field, I'll type something like: "All content for this project should be written in first person, under 300 words, and never use bullet points." That way, it applies to every single conversation in that project automatically. Claude saves that as a project memory, and now every new chat I open inside that same project picks it up automatically. And if I ever want to update or change those instructions, I can always go back and edit the project description or just update it directly in the chat.
The global memory works the same way but applies everywhere. So, anything I tell Claude about myself in any conversation gets carried forward to all future ones. And one thing worth knowing is that Claude won't just silently store everything. If I go to settings, capabilities, then memory, I can actually see every single thing Claude has remembered about me listed out. I can delete individual memories if something is wrong or outdated, and I can add new ones manually, too. So, if Claude keeps formatting things in a way I don't like, I can go in and add a memory that says, "Never use headers in short responses," and it picks that up immediately across every conversation.
I'll show you what this looks like. I'll open a new chat and ask Claude to write me a short LinkedIn post. Me saying anything about my tone or my style, it already knows how I write because it picked that up from our previous conversations. Without this turned on, every single conversation starts from zero, and you're wasting time re-explaining the same things over and over. The memory system [music] makes Claude smarter the longer I use it, but there's one thing it doesn't pick up on its own, and that's exactly how I want it to write.
Claude starts with a default writing style, and if you chat with it long enough, it does start picking up how you communicate, but that takes a while, and the results aren't always consistent. There's a way to set your style from the very first message. To do this, I'll hover over "use style" right here, and from there I can click "create and edit styles." [music]
The fastest way is to paste in examples of my own writing, like LinkedIn posts, emails, or drafts that I've written before. Claude analyzes them and builds a style profile based on my actual tone, sentence structure, and vocabulary. The other way is to describe the style I want in plain language, and Claude generates a profile from that description. I can create multiple styles and switch between them depending on what I'm working on. So, I might have one style for LinkedIn posts that's punchy and casual, another one for client emails that's professional but warm, and a third one for internal docs that's more detailed and technical. I just select the one I need before I start typing, and Claude adjusts [music] instantly.
Once the style is set, every response Claude gives matches it automatically, and I don't have to remind it in every prompt. I can just set the style right here, and it'll know exactly what tone I want. I'll show you the difference. I'll type the same prompt twice, once with the default style and once with my custom style. The default version reads fine, but it sounds like AI wrote it. But the custom version sounds like I wrote it. The sentence length is different, the word choices are different, and the overall feel matches how I actually communicate. If you're using Claude for any kind of writing, whether that's emails, social media posts, reports, or anything academic, this is the feature that stops you from having to rewrite everything Claude gives you.
So, now Claude remembers who I am and writes in my voice, but everything so far has been text. The next feature is where Claude starts building things I can actually interact with. Artifacts let Claude build things you can actually use right inside the chat. So, instead of getting a text response, I get a working calculator, a project tracker, a weekly planner, a dashboard with real data, or even a small game, all of it interactive and clickable right there in the browser next to the conversation.
And there are settings for this that a lot of users haven't turned on. I'll go to settings, then capabilities, scroll down to the visual section, and here I can see three options. And then I'm going to turn on artifacts. Now, there's also something a lot of people miss, which is the artifacts tab on the left sidebar. If I click on that, I can see a collection of pre-made artifacts that I can try right now without writing a single prompt. There are writing editors, brainstorm idea generators, raw note transformers, and a bunch more. I can click any of them, and it loads right into the chat ready to use.
But, I also want to build one from scratch. I'll type: "Build me a weekly content calendar where I can add posts, drag them between days, and color code them by platform." And Claude builds it right there in the chat. I can see the calendar, I can add posts to specific platforms and days, and drag them around. This would have taken me an hour to set up in a spreadsheet, and Claude did it in about 30 seconds. And the artifacts aren't locked inside the chat, either. Every artifact has a download button, and depending on what it is, I can export it as an HTML file or a standalone app I can host. So, if I build a client intake form or a pricing calculator, I can actually use it outside of Claude on a real website. And if I want to change anything, I just describe what I want different, and Claude updates the artifact live. That loop of describe, build, test, and adjust all happens inside one conversation.
Artifacts are built inside Claude, but all the information I work with lives outside of it in my email, my calendar, my drive, my Notion, and the next feature bridges that gap. Connectors plug Claude directly into the apps I use every day, things like Gmail, Google Drive, Google Calendar, Notion, Figma, Slack, and even PayPal. To set these up, I'll go to settings, then connectors, and then press "browse connectors." And when this page opens, I can see three tabs on the left: skills, connectors, and plugins. In the connectors tab, I can add Gmail so Claude can read my emails and draft responses. I can add calendar so it can check my schedule. And I can add Google Drive so it can pull documents into a conversation.
Now, the plugins tab is the part that almost nobody finds. Anthropic has built a collection of plugins that add specific capabilities to Claude. There's a PDF viewer that lets me view, annotate, and sign PDFs right inside the chat using a `/view/pdf` command. There's also a brand voice plugin. Each one comes with its own `/` commands that I can use in any conversation. To add one, download the Claude desktop app first. Once installed, open the app and navigate to the plugin page again, and that's where you'll be able to add it. I'll install the PDF viewer right now. And if I ever want to disable it, I can click into the plugin settings and turn it off from the top right.
At this point, Claude has my memory, my writing style, can build interactive tools, and is plugged into every app I use. The only limitation left is that I have to be sitting at my computer for any of it to work, and that's exactly what the last feature removes. [music]
This is the one feature in Claude that feels the most like the future, and it's called dispatch. Dispatch lets me trigger co-work tasks on my desktop from my phone. To turn it on, I'll go to settings, then co-work, and turn on dispatch. The way it works is simple. I leave my PC on at home, open the Claude mobile app on my phone, and send a task. Co-work executes it on my computer and sends the results back to my phone. Every skill and connector I've already set up carries over, so dispatch has access to everything co-work does.
So, if I'm at the office and I realize I forgot a file on my home computer, I can pull out my phone, tell co-work to find that file, and send it to me on Slack, and it does it. If I'm on my commute and I want my meeting prep document ready by the time I sit down at my desk, I dispatch it from the train, and it's waiting for me when I arrive. And the tasks aren't limited to simple file searches. I've dispatched full research tasks where co-work opens my browser, pulls information from multiple sources, compiles everything into a document, and saves it to my drive, all while I'm away from my desk. By the time I get home, the work is done and waiting for me.
The whole point of dispatch is that your AI assistant doesn't stop working just because you stepped away from your computer. As long as the desktop app is open and your machine is on, you can trigger anything from anywhere. And that dispatch feature alone changed how I use Claude because I went from only being able to use it at my desk to being able to trigger full tasks from my phone while I'm on the train, at the gym, or anywhere else.
If you want to see more videos like this where I break down the features and tools that actually make a difference, hit the subscribe button so you don't miss the next one. Hit the subscribe button so you don't miss out. Thank you for watching, and I'll see you in the next one.