#151 Perplexity Computer vs. Manus: Which One's Actually Better?

Show Notes

Corey puts two leading agentic AI tools head-to-head in a real-world showdown. He tests Perplexity Computer ($200/month) against Manus ($20/month) across three practical use cases: multi-source research with website deployment, connected app integrations with Google Workspace, and data collection with visualization. You'll see exactly how each tool performs on speed, accuracy, output quality, and which one earns Corey's recommendation for 2026.

Join the Build With AI community waitlist: https://return-my-time.kit.com/1bd2720397

• Perplexity Computer completed tasks 5-10x faster than Manus across most tests — finishing in 60 seconds what took Manus 10 minutes

• Both tools nailed data accuracy when researching AI startup funding, pulling identical figures from TechCrunch

• Perplexity's Google Workspace integration produced cleaner, properly formatted deliverables (tables, headers, structured docs) while Manus output jumbled text

• For website deployment, Manus actually had slightly better UI design — but both successfully created shareable public URLs

• When stuck, Perplexity self-corrected and found the data; it recovered gracefully from errors

• ClickUp was recommended by both tools as best value for project management pricing ($12/user/month for Business plan)

• Perplexity Computer wins overall for speed, output quality, and cleaner UI despite the 10x higher price tag

Timestamps:

00:00 - Introduction and test overview

00:30 - Pricing comparison ($200 vs $20 plans)

01:00 - Test 1: Multi-source research + website deployment

14:04 - Test 1 results: Speed comparison (2 min vs 10 min)

16:25 - Website design comparison

18:30 - Test 2: Gmail + Calendar + Google Docs integration

27:14 - Test 2 results: Output quality comparison

29:40 - Perplexity's formatted table vs Manus' jumbled text 32:00 - Test 3: Data collection, analysis, and visualization 39:30 - Test 3 results: Bar charts and pricing spreadsheets 41:52 - Final verdict and recommendation

Links Mentioned:

• Perplexity Computer vs Claude Cowork: https://youtu.be/_s0ymvl6vNc

• Perplexity Computer: https://perplexity.ai/

• Manus: https://manus.im/

What is Build With AI?

Most AI podcasts talk about what's possible. Build With AI shows you how it's done, live. Each episode, host Corey Ganim brings on entrepreneurs and operators who share their screen and build real AI automations, workflows, and tool setups right in front of you. No boring slides. Nothing that hasn't been battle-tested. You'll watch actual implementations get built from scratch so you can follow along and do the same in your business. If you're a non-technical entrepreneur who wants to put AI to work without becoming a developer, hit play and build along with us.

Corey Ganim: All right. So today we're going to be putting perplexity computer head to head against Manus. Now I love both of these tools. Manus was actually the first agentic tool that I ever used probably eight or nine months ago. Uh, but perplexity is my current favorite tool, uh, specifically perplexity computer. So I'm going to share my screen and show you guys exactly the three tests that we're going to be putting these tools up against now for what it's worth. I'm on the perplexity max plan, which costs $200 a month. And then I'm on the Manus pro plan, is their $20 a month plan. Now it shouldn't make a huge difference as far as the output quality, but just wanted to put that caveat in there. So the first test project that we're going to put these two up against each other with is we're going to do. Multi-source research and then we're going to have them deploy that research to a website. So the test that we're going to run is can each tool research real data from the web, generate working code and deploy a live functional website, all from a single prompt. And this is the prompt that we're going to be using here. The second test we're going to run is can each tool read data from a connected app, which in this case will be our Google email calendar drive, and then cross-reference it with web search and then produce a deliverable all without manual copy and paste. And then the last test we're going to run is can each tool browse multiple live websites, extract structured data, analyze it and produce the polished chart and summary all fully autonomously. Now, I chose these three case studies because they're case studies that you cannot, you cannot do these tasks with a normal chat bot, like a chat GBT or like a Claude. So I wanted to make these use cases specific to agentic tools like perplexity computer and Manus. So with that said, let's go ahead and throw the first prompt into each tool. So we're going to copy and paste it directly into Manus. We're going to click go and then we're going to copy and paste it directly into perplexity computer and press go. And then I will come back shortly once both of these tools are done. All right, we are back now. This task took about ⁓ nine to 10 minutes in total, but what's interesting is perplexity finished in about two minutes. Manus took about 10 minutes for the exact same task. So what we're going to do first is we're going to look at the live summary of the data that we asked it to produce. And we're going to compare that against the real information that it pulled it from. So. Perplexity says the five most funded AI startups of 2025 are open AI at 40 billion and Thropic at 16.5 scale AI at 14.3 project Prometheus at 6.2 and X AI at 5.3 billion. And it says that it pulled this information from tech crunch. So let's look at the actual tech crunch article. So tech crunch says open AI at 40 scale AI at 14.3 and Thropic at 13 Prometheus at 6.2 and X AI at 5.3. So It looks like it nailed the figures exactly except perplexity has Anthropic at 16.5 million and Anthropic according to TechCrunch was only 13 billion, but it turns out I think perplexity pulled that data from both their E funding round and their F funding round, series E plus series F. Whereas the TechCrunch article is only reporting on the series F. So theoretically perplexity is still right. We asked it to deploy, ⁓ basically turn this information into a website that has a publicly shareable URL. So this is a more subjective rating, but if we look at the website that perplexity created and keep in mind, it built this in like two minutes. This is the website. So pretty simple, pretty cut and dry here. It's got a leaderboard as far as the funding is concerned with, ⁓ open AI leading as we said, you know, nothing too fancy here, right? This looks pretty good for a simple. One-shot exercise, we gave it no guidance or feedback in terms of UI design or how the website should look. So all things considered, I mean, the fact that we gave it no guidance, this looks pretty good. And it is shareable in the sense that in the top right, we can click share. Right now it's private only to me, but I can change it to where anyone with the link can ⁓ view this website so that it passed the test of creating a website with a publicly shareable link. Now, if we go over and look at Manus. So Manus again did the exact same thing, but took it about 10 minutes total. It found the exact same results. Open AI at 40 billion, Anthropic at 16.5. It also, which is interesting to note, combined Anthropics series E with their series F into 16.5. Scale AI at 14.3, Project Prometheus at 6.2 and XAI at 5.3. So as far as the accuracy of the data, it is spot on. Now here's the actual website that Manus built. So we see it's a little bit of a different theme, more of kind of like a spacey vibe, white and blue font, and then a little bit of gold thrown in as well. It laid it out very similarly to how perplexity laid it out. It's just a different UI design overall. So again, this is a very subjective rating as far as like which website design I prefer. I think the Manus website design actually looks a little better. I just like this. style and this vibe a lot more. But as far as accuracy of the data and everything that we asked it to do and deploying it to a public URL, both of them completed the task perfectly in an acceptable fashion. Again, in the top right, we can click publish to where anyone with the link can view the website. So it passed that as well. So all in all, I'd say this first task was a tie. They both did exactly what we asked them to do. But as far as the UI design, think Manus slightly edges out perplexity. computer in this first exercise. All right. So now we're going to do, ⁓ we're going to give it a new task. We're going to give it the second task in the use case test here, which is to check my Gmail for the most recent email. And then, which is a test email that I sent it. It's going to extract project names and deadlines from it. Then it's going to check my Google calendar for any conflicts with those deadlines in the next 30 days. And then lastly, it's going to create a Google doc titled. Q2 planning summary that includes a table with each project, its deadline, and then whether there's a calendar conflict on that date. So as we did before, we're going to go back into Perplexity Computer, hit New Task, paste in our prompt, click Go. Same thing with Manus, paste in our prompt, click Go. And I will be back when each of these are done. both tools have now completed the task. Now, again, it's worth noting that, ⁓ Manus took again, roughly 10 minutes to do this task, but I think is a relatively simple task perplexity computer finished it in about 60 seconds. So if speed is a priority perplexity computer thus far is absolutely the winner in that category. Now let's check out the final result. So again, what we wanted it to do was check our calendar. or first of all, extract a specific email that I had sent to myself with our Q2 planning, ⁓ project names and deadlines, then check my calendar for any conflicts with those deadlines. And lastly, create a Google doc that includes a table with each project, its deadline, and whether there's a calendar conflict on that date. So let's check out what we've got here. So Manus outputted the direct link to the Google doc, which we have right here. Now, so. ⁓ What it gave us is a pretty basic Google doc. It looks like they tried to put it into a, ⁓ table here because it has like the column, what, what looked like our column headlines here kind of stuff together in the document, but it's not in a calendar format. or sorry, it's not in a table format. And if we look at the prompt that we gave it, we specifically said include a table with each project, its deadline, and whether there's a calendar conflict on that date. So it did what we asked it to do, but it didn't output it in a table. just outputted it in a bunch of, ⁓ kind of jumbled. It just outputted it in a jumbled format that kind of makes it hard to read. So did it do the tasks that we asked it to do? Yes. Did it take a long time to do it? Yes. And did it give us in the exact format that we asked it to give it, give it to us in? ⁓ not exactly. So. Man, it's didn't do great with this task and it took a long time to do it. So let's go to perplexity computer. And see how it performs. Now, keep in mind, perplexity did this in roughly 60 seconds. And what's interesting to note is I actually don't have my Google docs directly connected to perplexity computer or Manus, but it gives it to you in a doc X that as long as you have Google drive connected, which I do to both tools, you can export it directly as a Google doc to your Google drive. So that's kind of a little nuance that as long as you have drive connected, you can utilize. Google's entire ecosystem. So let's do that. Let's export perplexities deliverable straight to Google drive and see it in a Google doc format. So it's a one click and here we have it. So it looks like perplexity computer did a great job with this task. So it did exactly what we asked it to do. It put the projects, the deadlines and the calendar conflicts into a table, which Manus did not do. This is a good clean looking table. ⁓ it even kind of put headers onto the document as far as like, here's the source email. Here's a section for projects, deadlines and calendar conflicts, and then notes at the bottom. So when you compare this output to the output that Manus gave us, mean, perplexity absolutely blows it away. Like Manus didn't even format the deliverable Google doc at all. It kind of just threw a bunch of texts into a Google doc, whereas perplexity computer. Went a step further, formatted it properly, pulled the source email contents, put it into a table, just like we asked and then formatted the notes as well. And something I was curious about whether or not it would do, it did keep it to one page, right? We'd specifically said, make it one page. Both of those tools did that, but for this particular exercise, perplexity computer absolutely blows mannus out of the water and definitely takes the cake for the second use case. Now let's go in and give it the final. Use case here, which is data collection, analysis, and visualization, where we're going to tell each tool to go to the website of five popular project management tools. In this case, Asana, monday.com, ClickUp, Notion, and Linear. And for each tool, it's going to find the price of their most popular paid plan. And then it's going to compile the pricing into a spreadsheet. And then it's going to create a bar chart comparing all five tools by price. And then it's going to write a short summary, three to four sentences each. Identifying which tool is the best value and why. And then lastly, it's going to deliver the spreadsheet, the chart image and the summary as separate files. So as before, we're going to go into perplexity computer, simply hit new task, copy in our prompt, press go. And we're going to do the exact same thing in Manus new task, copy in our prompt, press go. And I'll be back with the output of each tool. So they are both done now. Let's look at the output first starting with Manus. So remember we asked for three deliverables from each tool. One, we asked for a spreadsheet with all the pricing data and the source URLs. Next we asked for a bar chart, ⁓ basically comparing all five tools visually. And then lastly, we asked for a summary, no more than four sentences ⁓ with the recommendation for which tool would provide the best value. So from Manus, here's the bar chart that it gave us. Pretty simple image here. We see the best value box kind of overlaps with the pricing here. So it's like a slight visual flaw, but all in all, pretty good looking bar chart, no complaints there. Next, we're going to look at the spreadsheet that it gave us comparing the pricing. So it did exactly what we asked. It took Asana, monday.com, ClickUpNotion and Linear. And I did verify this pricing. So this pricing is all accurate. We did ask it to give us a quote based on the annual pricing option per seat. So this is all accurate and it included the source URLs. So it did that process just fine. And then lastly, we asked for a no more than four sentence summary of which tool it recommends. So I'm really just going to count the sentences here. So that's one, two, three sentences. So this did exactly what we asked it to do and Manus. Recommended click up as being the best overall value with the business plan that it offered. So another thing to notice, Manus actually did this task a lot faster than perplexity computer perplexity computer actually got stuck for a little bit trying to find a sauna's pricing tier. And it took a couple of minutes, but it did fix itself and it did find the information. So this is the one task that Manus beat perplexity in terms of speed. But now let's look at the output from perplexity. So. The first thing that it gave us is the ⁓ spreadsheet that we asked for listing the pricing and it pretty much gave us the exact same output as Manus. All the pricing matches what Manus gave us with the source URLs, annual cost per user, price per user per month billed annually. So it did exactly what we asked it to do there. Next, we asked it for a bar chart comparing the pricing of each tool. Now I will say perplexities bar chart is a little better. Again, this is subjective, just my subjective opinion, but there's no overlapping in terms of like the font or the numbers or anything like that. Pretty clear and easy to read. So perplexity maybe edges it out slightly just on for that deliverable. And then lastly, the, asked it to generate a summary listing its recommendation for the tool that we should use. And it looks like perplexity also recommended ClickUp's business plan. And let's see how many sentences it gave us. So it gave us one, two, three, four. So it also did what we asked it to do and gave us four sentences or less. And it also included the sources in the pricing summary, which we didn't necessarily ask it to do, but it's kind of a nice touch. So for this final task, it really is a tie. mean, the output is roughly the same across the board. ⁓ Manus did just a good of a job here, even a little quicker than perplexity computer. So you're probably wondering at this point, what is the consensus? What is my overall recommendation between Manus and perplexity computer? And considering the fact that perplexity ⁓ blew Manus out of the water for the second task ⁓ the fact that it's five to 10 times faster and it's output, I give perplexity computer the edge over Manus. I think the UI is cleaner. think it's easier to use. I think it's more powerful overall, and it would definitely be my recommendation if you're deciding between Manus or perplexity computer. So I hope this comparison video is helpful. I have another video comparing perplexity computer head to head against Claude cowork. We will link that in the description below and the show notes. Be sure to go check that out. And if you like comparison videos like this, then leave a comment or send me an email and let me know. And I'm happy to do more just like them. So thanks for. watching and I'll be back next week.