Manus AI: Best Uses for Design, Research, and Slides (Full Transcript)

[00:00:00] Speaker 1: Move over, ChatGPT, Manus AI just stole your job, or at least part of it. It's an AI agent that claims to do a lot, but it does get expensive really fast. So here are three things you'll actually wanna use it for, starting with Design View. I'm David, working with Kevin to bring you more great tutorials like this one. To follow along with me, head to the link in the description below to get 500 free credits for your free trial of Manus. Now, if you've used ChatGPT before, you already know how to use Manus. When you sign in, you land on this homepage. You'll see your past chats in the left sidebar, and of course, the large prompt box in the center. That's intentional. Now, here's what's different. Along the bottom of the prompt box, you'll see specific modes you can use. And in my opinion, these are where Manus shines compared to the foundation models. And as expected in 2026, you can connect your own apps to Manus. This includes popular apps like Google Workspace and Microsoft 365, as well as custom MCP servers, and also custom APIs, which is a nice touch. You'll also notice a number of pre-built prompts directly on the homepage. If you're struggling to figure out how to make use of Manus, these can be great inspiration. You can click on any one of them to populate the prompt box, then just fill in the placeholders, and off you go. Now, I promised you three tasks that Manus steals from ChatGPT. So let's start with number one, Design View, by clicking Design in the mode list. If AI image tools have ever wasted your time, this fixes that. Notice in Image Mode that we are using Nano Banana Pro. This is Google's premier image generation model. You can also downgrade to ChatGPT's Image One model. Now, this is important to understand about Manus. It is not its own model. Instead, it orchestrates models like Nano Banana Pro, models from Anthropic, and even Alibaba. Now, if you've watched me on this channel before, you know I have a donut shop called David's Donut Den. I wanna make an image of this donut shop that I can use for a landing page. So with Design View, I'll prompt it to make this large photorealistic image of the donut shop and give it some restrictions about how I'll use it on that landing page. Let's see what it does. It takes less than a minute and comes up with a nice image given the constraints that I gave it. Here's where Manus wins against ChatGPT. Notice the right-hand side of the page is not just an image viewer, it's a canvas. And when you click on the image, you get built-in editor tools. Manus' real differentiator is in this edit experience. The first tool to show you is the comment tool. So if you're following along, click on this comment button down below. It can be a real pain to make changes to a specific part of the image with AI. But with the comment tool, you can click on that part of the image you wanna change, then add a prompt, like add a couple in the picture, and then apply the changes. You can also make multiple comments at once and apply all those changes at once too. Again, after 30 seconds to a minute, you'll get the result. Now, the nice thing is if you don't like it, you can always delete it. To do that, just make sure you select the photo with the select tool. The other nice thing is that you still have the original image to work with. In fact, let's take advantage of that and use the edit text tool. This identifies text areas that you can edit. So if for some bizarre reason, I wanted to rebrand David's Donut Den, I can totally do it. So let's see the damage with Dave's Donut Den. Again, you can batch up text edits before you click apply, but since we only have one, let's see what happens. Okay, that's not too shabby, but don't start calling me Dave in the comments. One last thing to show you here is how to smartly combine images. Here's what I mean. You can upload your own image into the canvas by clicking the center button. Let's add this old David's Donut Den logo. Now here's the fun part. Let's scroll up a bit to get some room. Now click the dropdown next to the comment button and select the mark tool. I wanna add this logo to the door. So first I click on the logo. You see how it sent it directly to the prompt box. Now I'll click and drag on the door window where I want that logo to appear. And then I can add some text to the prompt to describe what I want done. This should totally resonate with you if you've ever beaten your head against a wall trying to edit an AI image. And I put my foot in my mouth there because this is what we get. It's good this happened because this illustrates the importance of your prompt. My mistake was saying storefront window, not the door window. Fortunately, I still have the original images here so I can fix it. And after repeating the process with a smarter prompt, the logo is successfully on the door. Making new images for my donut shop is exciting, but behind every successful new product is seamless planning, financial tracking, and inventory management. That's why for my real business, I rely on the sponsor of today's video, QuickBooks Online. With it, I can keep track of our products, invoice our customers, monitor sales, and track our growth. Plus they've got AI agents to save me even more time by auto-categorizing transactions and sending custom payment reminders. With all my financial data in one place, QuickBooks helps me make smarter decisions and gives me the confidence that I'm on track. Check out the link in the description below to get a discount off your first three months of QuickBooks Online. Now, let's address the biggest complaint about Manus, the price. Manus runs on a credit system and you will be confused at first and I'm here to help. Credits are payments for AI usage. And here's what that means. The image work you just saw cost about 100 credits, which is about 50 cents US. You can get a free trial for seven days, which gives you 8,300 credits. And that's more if you use our link below. So the rest of this video answers one thing. Even if Manus is better, is it actually worth paying for? If you've wanted ChatGPT to research a list of things, you've probably been disappointed. In fact, you'll probably hit a wall around 20 to 30 things to research at once. That's where Manus' wide research mode comes in. To access it, click the More button in the bottom right and then click on Wide Research. You'll wanna use this for big research projects. For example, I wanna resell my donuts across the Pacific Northwest. So let's find 200 coffee shop owners, their contact information, and draft an initial outreach message I can send them about a partnership. Here's what happens when you submit this prompt. First, it's going to scour the internet to try and find these coffee shop owners, but it won't do it one at a time like ChatGPT. Instead, Manus spins up sub-agents to process each contact. So if you have 20 sub-agents running at once, that means 20 contacts processed at once, not one. At the beginning, it uses a combination of its own search tools and a web browser to find the leads. After seven minutes, it's found 200 coffee shops, and now it gives us the wide research trigger. Now be careful because wide research can take up a ton of credits. So hopefully you can make money with the research that you're doing. And honestly, that's what makes it perfect for lead research. Now I finished this task a bit earlier. It took about 45 minutes to complete, and you can see we got a decent amount of contact information from it. We also got our outreach messages and several lead lists. These are all available for download at the end of a task. I compared this with ChatGPT 5.2 Thinking and 5.2 Pro. The Thinking model basically said, I'm not gonna do this, but here's how you can do it, which obviously Manus beats. Now ChatGPT Pro finished faster in about 27 minutes, but it also didn't find any owner contact information. Then I asked Claude Opus 4.5 to judge the results based on contact info accuracy. Manus came out on top. Now you may be wondering what that cost. It came out to 2,057 credits, which is just above $10. Assuming you're not doing too many of these per month, this is a bargain compared to the $200 a month you'll have to fork over to OpenAI to access their Pro model. Now out of interest, I did the same thing scraping for LinkedIn leads instead. That cost almost 4,800 credits, which is $24. It was also a huge pain because I had to keep logging in to LinkedIn to get it to do its job. In my opinion, there are better ways to get LinkedIn leads. Now, one thing I didn't mention is that for this exercise, I used the Manus Max model. This is like their Pro version and I found better research results using it. Now out of interest, I tried a similar prompt for 20 coffee shops instead of 200. And then I compared the results of Manus Max, Manus, ChatGPT Thinking, and ChatGPT Pro. And then I used Claude to judge not just the contact accuracy, but also the quality of the email outreach. Surprisingly, ChatGPT 5.2 Thinking came out on top. And then ChatGPT Pro, Manus, and Agent 4, which is Manus Max, ended up about the same. So for smaller research tasks around the 20 to 30 item level ChatGPT 5.2 Thinking is probably gonna give you more bang for your buck. Now all of these models make mistakes, so I wouldn't trust any of them 100% with a research task. But the bottom line is that at scale, Manus gives you more usable data for the price than the other models and hiring someone to do the research for you. Now we're all still waiting for AI that can actually build usable slide decks. So let's see what Manus can do. This is also a special mode which you can access by clicking Create Slides. You may have watched my GenSpark tutorial a few months ago where I asked for a slide deck. I'm gonna use the same prompt in Manus. Let's start with just this for now. No fancy prompts and no templates. It takes about seven minutes start to finish, and the first step is creating a slide outline. If you want more control over the output, provide this slide outline to Manus instead of letting it generate it. Then it creates and refines the slides one by one until it gives us the final deck. In my opinion, this is a really clean design across all of the slides. As I click through these slides, you'll see it's made up a lot to fill in the text. To edit any of this text, simply click on it and make your changes. Now let's say you find something that you don't quite like. Like this image doesn't really make sense for the Amazon gift card contest. Here's how to fix it. Click the magic wand icon at the bottom right of any slide, then add your prompt to suggest the change to the slide. It knows you're referring to this slide, so you don't have to put that in the prompt itself. And after about a minute, the image is replaced. And while you can run the slides directly from Manus with the play button here, you'll probably want to export this to share it. Click on the three dots here to export it to Google Slides or click on the download button to export it to Microsoft PowerPoint. Now that was pretty good, but here are the options that I skipped over. Scrolling down from the prompt section, you'll see templates. I recommend you start with no template or with Manus' built-in templates. The built-in templates will dominate the feel of the presentation. For example, here's what the whiteboard template looks like. And this is the sketch template. Here's a template I exported out of Canva. It had a lot of images in it that Canva wasn't so good about laying out on the page. So my recommendation is if you have images that you want to put in the presentation, you add them after you export it. Finally, don't miss that you can change the number of slides that you generate with this dropdown. So now you've seen where Manus AI could be a good fit for you. Design view if you need more image editing precision than Chat GPT or Nano Banana Pro, wide research to research hundreds of items or leads, and slides to create stunning visuals around a topic. Before you go, check out the QuickBooks Online deal in the description below. I'm David, and I'll see you in the next video.