ChatLLM Operator | New 1 Click AI Agent automates your work and research

--- type: transcript date: 2025-04-01 youtube_id: Y1266iiwOJg --- # Transcript: ChatLLM Operator | New 1 Click AI Agent automates your work and research In this video, I'm going to be showing you Chat LM operator. It's been a while that we've been able to write and brainstorm and even translate things with large language models. But what chatm operator allows you to do is not just write tasks, but effectively use a computer like you do. A number of months ago, Chat GPT released their version of operator. But one of the reasons that I don't think it really took off is its $200 a month price tag. One of the really impressive aspects of Chat LLM operator is it's just 10 bucks a month. The way that you can leverage this is you can put in your query similar to how you would hand off a task to another person. You could say I want to book a flight to a particular destination. Here are some requirements. And what it will do is it's going to autonomously go through and be able to control the mouse on the screen, be able to use the keyboard just like you would. And how this works is there's essentially a loop occurring where it's taking screenshots. It's passing those screenshots to a large language model and then based on the response from the language model is it's going to give further instructions as we see here with just a simple instruction. Show me all of the available flights to Greece. And and what it's going to do here is it's going to stream out all of the actions that are occurring on the left hand side here. And the way that you can think about it is essentially a computer that you can access in the cloud. The really cool thing with this is you're able to sign in and authenticate into your accounts and it will securely store all your session tokens and what have you for all of the various sites that you use. The past couple years has really been about large language models processing a number offormational tasks. And I think in a lot of ways you can think about this as getting ahead of the curve in terms of how a lot of AI agents are going to navigate the web. Fundamentally, the internet was not built for AI agents. Right now, there's some initiatives around MCP and that new protocol, but it's going to be quite a while before we have a standardized format for how agents interact with the web. What those agents are going to have to rely on are things like this where you have this agent that can navigate the web and perform different tasks within different applications that you have. The one thing to know with this virtual machine is you can access it all directly from the web browser. So, at any point if you do have to intervene, like here it's telling us to put in our email address, continue on, add in our password and what have you, you can go ahead and click on the screen just like you would a typical browser. I'm going to go ahead and authenticate here. Now that I'm authenticated, I can say I am now logged in. And the nice thing with this is it's really just a collaborative process back and forth. And I think the thing with this is it really is starting to get you familiar with the idea on how you can leverage these AI agents. What I could do within here is I'm going to say I want to go to the Hacker News website and I want to log the top five stories in a brand new spreadsheet. One of the use cases that I've been thinking about for these types of agents where they can be useful is when you're trying to get data from one place and then have that data move to another place. This is a really quick example where it went over to Hacker News. It went and it found those top stories and now it's navigating over to Google Sheets. It's going to create that sheet for me and add in those top stories. Here the agent goes. It's clicking each of them and then it's putting in the respective information. From there we see it's putting in the rank, the title as well as the URL. And the thing with this type of system is this is the type of system that is only going to improve as these large language models that are at the core of it begin to improve. But already, you would probably be surprised at the number of different tasks that you're able to do with this. With just very simple instructions, similar to how I would ask someone to do a task like this, it was able to go through, navigate the web, find the correct information, parse that within a spreadsheet for me. Now, I want to demonstrate a bit more of a complicated task. Next, what I'm going to say is I want you to go through my email, see anything that's unread and anything that needs to be actioned. If there's anything that requires me to schedule any time, make sure to check my calendar first before sending a reply. And then finally, once you're done, go and update everything that you've done within a new Google Doc. And let's call it Friday's operator report. I'll go ahead and send this through. Now, obviously, this is a little bit more of an involved task because there are a lot of substeps as well as specific instructions that I instructed it to do. I want to make sure it checks my calendar. I want it to go and read the unread emails. Here we see it's opened up the latest unread email where we see project alpha. Can we book a couple hours to go over some time? And now it's going to go back to the inbox. It's opened up our calendar and it's went over to actually search for next week, which is when the time was requested for. Now it says, I can see that Monday afternoon appears to be free. And what it's going to do here is we're going to go over to that message that we had just taken a look at. And here it's going back into that message. And what it's going to do is it's going to draft a reply with the specific time that it was able to find on my calendar. And there we go. It went and it's sending that message for us. Now it's going over to the next email. And similar to the last process is it's going to go over to our Google calendar now that it's already open and it's going to go and send and write that reply for us. Now it sees an item that it doesn't need to action. And finally, we see it went over to Google Docs. it wrote out the title for us and then there are all of the different actions that it took for us. The way that you can think about this is there's going to be specific workflows that you take dayto-day whether it's repetitive browserbased tasks that you dread doing. It could be compiling daily reports. All in all, the way that you're going to be leveraging Chat LLM is going to be largely different than everyone else. There might be different tasks that I care about that you care about. And maybe you have a browserbased task that you absolutely dread doing. It could be compiling daily reports, maybe checking multiple sites for various updates. It could even be filling out forms over and over, or managing different tasks that you have in mind. The real power of tools like Chat LLM Operator really come alive when you apply them to your own pain points. While I showed a few different examples like data scraping to spreadsheets and email management, the potential of this type of application are increasingly broad. I'd encourage you think of something that you want to automate with an AI agent like this. And I'd be interested to hear all of the different use cases on how an agent like this can be useful for automating various tasks in your day-to-day process. Overall, that's pretty much it for this video. If you found this video useful, please comment, share, and subscribe. Otherwise, until the next