
Sign-up for the Warmwind waitlist today: https://warmwind.space/?src=L20&md=yt Streamline Your Work with Warmwind: The Ultimate Autonomous AI Platform Learn more about Warmwind here:@warmwind_OS In this video, we explore Warmwind, a platform allowing you to create and deploy autonomous AI agents using natural language. The demo includes automated tasks utilizing tools like ChatGPT, Gmail, Midjourney, and Firefox. Learn how to equip workspaces with different applications and automate repetitive workflows. The video also showcases the teaching mode, where you can train AI agents to understand specific tasks. Sign up for the waitlist to access this powerful platform. 00:00 Introduction to Warmwind 00:07 Platform Overview 00:09 Setting Up Applications 00:36 Automating Tasks with ChatGPT 01:31 Creating AI News Digest 02:36 Advanced Workflow Automation 05:30 Teaching AI Agents 06:33 Access and Final Thoughts
--- type: transcript date: 2025-07-03 youtube_id: hHfuCHH51xs --- # Transcript: Warmwind: Autonomous Cloud Employees in 7 Minutes In this video, I'm going to be showing you Warmwind, which is a platform that allows you to create and deploy autonomous AI agents with just natural language. First up, just a brief overview of the platform. Within here, I have a number of installed applications at the top of the screen. Chat GBT, Gmail, some other Google products, as well as Midjourney installed. We have a number of different applications that we can leverage. We have things like Amazon Chat GBT. We can even equip it with a Chrome browser or something like Firefox. The way that this works is you're going to be able to equip the different workspaces with particular applications that you want to leverage. From there, you can begin to automate different tasks that you have. First, what I want to quickly demonstrate is I'm going to say, let's use chat GPT to research the latest AI developments and summarize key points. Once we have that, send that to me at developers.digest.ai@gmail.com. The really cool thing with this is what you'll see at the top of the screen here. With one simple click, you can add in different applications within this. In this space, I've equipped it with Gmail, a number of other Google products, Midjourney, as well as Firefox. In essence, since it has access to a web browser, you can leverage this or teach it different workflows to accomplish effectively any task autonomously behind the scenes. Within here, we see that it's opened up Chat GBT and all of the different interactions that you see on the screen here, whether it's the cursor moving or it typing out into input boxes, all of that is actually the AI behind the scenes. It asks ChatGpt to put this within a document. And the really cool thing with this is because chat GPT does have the ability to actually create these documents. We've copied that structured document that we've generated from Chat GPT. We can see that it's opened up a new Google doc here. And within here, we have the structured document that it just generated for us. We have AI news digest week of July 1st. And then we have the nice document all broken out with all of the key information. As soon as we have that, we see that it's selecting the title of the Google doc. And then from there, I can see that it's adding in my email. So, it's also attaching a message of here is our AI report. Now, if I go over to my email, I have this document that it generated from chat GPT, put it into Google Docs, and then sent me that document where I can iterate on that. The really cool thing with this is I could also set and instruct Wormwind to say every day or every week, send me the latest in AI news. Put it in a Google doc. It's in an editable format. And what you can have it do is behind the scenes what I just showed you. It can do that without you having to actually monitor it. Think about the different workflows of repetitive things that you do every week, every day, whatever it might be. And you can begin to offload all of these different workflows to the agents that are within Wormwind. Now, I'm going to demonstrate one more query. I'm going to say, use Firefox to identify trending topics on hacker news and GitHub trending. Then select the top two most interesting AI or development related topics. Let's use chatbt to write a blog post about these trends. Midjourney to generate a matching visual and Google Docs to compile everything. Then Gmail to send me the publish ready content. I'm going to send this through just to show you how it can accomplish much more complex workflows as well. What we'll see within here is since now this task involves leveraging a web browser, what it's going to do in this case is we're going to be leveraging Firefox within here. It can interact with the web browser just like you would. We can see within the URL bar, it's putting in news. Combinator.com. It's loading up the Y combinator site. And then what it's going to be able to do is have all of the context of everything that we have on this page here. We see all of these relevant stories. And what we've asked the agent to do is filter it based on AI news as well as development news. I can see it's scrolling through the page here. And once it's determined that it's found what it needs to do, we'll see the agent continue. So now that we have the information from Hacker News, I can see within here it's opened up chat GPT. ChatGpt has wrote out this document for us. And similar to the first time, we're copying that document. In this case, we're going to open up Google Docs, create a new document, paste in what we had just generated from chat GPT. And then as a last step, we're going to be going over to MidJourney to generate a matching image and visual and put that within the Google doc and compile everything. Then we're going to be leveraging Gmail to send me this document. similar to the first time, but this time we're leveraging some other tools. Now, I can see that it's opened up MidJourney for us. We're going into the prompt window here. It's gone ahead and put in a prompt to generate something from Midjourney that navigated over to the create tab. It's adding and uploading those images that it generated from MidJourney into our blog post Google Doc here. And you can begin to get a sense on just how powerful it is to have these autonomous AI agents just work behind the scenes. And you begin to get a sense on how you can leverage this. Now it's going into Gmail again and it's adding in that document just like we had specified. Now just to show you some other aspects of the platforms within here you can equip your different workspaces with all of these different applications. You have things like Amazon, Canva, Chrome, Duck.Go, email, Google Calendar, so on and so forth. Within here we see all of our installed applications here. But what's really neat with this is you can set it up to have different workspaces. For instance, I have Astro's workspace here. Let's say for instance I wanted to equip this workspace with these particular applications but I wanted to create another workspace that leverages a whole host of other applications. I can create a brand new workspace and I can create this with a fresh set of applications that I want to use. Maybe I want to have something that leverages eBay as well as Gmail. And then similar to the prompts that I showed you, you can go and instruct from there what you want to do. Now, one really cool thing as well that you can do is you can actually teach the AI agents. Let's say for instance, I want to leverage something within Chrome and there's a particular workflow that I want the agent to understand. The cool thing with the teaching mode is you can teach it similar to how you would teach another person. I can describe within my microphone all of the different things that are occurring, any relevant context, and then I can go through and select the different workflows, whether it's within Chrome or through the different applications. And as you go through, effectively, you're teaching that AI agent to know the workflow. And then within here, I can actually just instruct whatever I want the AI agent to do. Let's say I have a CMS that might not be within the training data of how an LLM could navigate. Well, what I could do is I could go into Chrome. I could select to upload documents or whatever the particular workflow is. And then as soon as I instruct the agent again within the workspace, it will have the context of how to actually perform those things. It is really both intuitive and clever in terms of how they've set this up. Now, in terms of accessing Wormwind, right now, you will have to sign up to their weight list. Within here, you can also see a ton of other examples on how you can potentially leverage Warmwind, but otherwise, I just wanted to do a really quick one showing you the platform. Kudos to the team over at Wormwind for what they built. It's an incredibly powerful platform. But otherwise, if you found this video useful, please like, comment, share, and subscribe.
Weekly deep dives on AI agents, coding tools, and building with LLMs - delivered to your inbox.
Free forever. No spam.
Subscribe FreeNew tutorials, open-source projects, and deep dives on coding agents - delivered weekly.
Technical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.