Thank you so much @Wyndo for giving me this opportunity to write about it.. It was so much fun researching, playing and writing about this comparison of Manus and Claude Code. I am still staying with Claude Code but it was so easy to use and setup AI Agents with Manus, it was pretty amazing..of course the the credits costs :)
Thank you for this test! For me, these are all use-cases where I use Perplexity. So basically all that needs web research in some form, I give to Perplexity (tasks, computer etc)
Especially some of my personal and health related things (this is why I'm looking forward to the new Health features in PPX with connectors to Oura etc). I recently started a new med for a chronic illness that also affects my sleep and I'm doing a daily diary with my Whoop and Oura data, my subjective experience and Computer is not only writing this in a diary, but also giving me the context based on pharmacology - basically being my doctor in my pocket.
But also for business stuff and research, I use PPX (with of course Claude as base model). So currently I haven't found a use-case where I would need OpenClaw or Manus.
Yeah i use perplexity computer too. Will have another deep dive for it in the upcoming post. But perplexity is kinda costly. For agentic task, it can spend hundreds of credits. But i kike that it can automate almost anything and have many app integration
Right, especially since you don’t get credits while being on the Pro plan - and the Max plan is totally unreasonable priced, especially when you got a Max plan already in Claude.
The 'control vs simplicity' split is real and I keep running into it. I use Claude Code for structured pipelines where I want to know exactly what the agent touched and why. The precision is worth the setup cost.
What surprised me about your breakdown: the 5-10x cost gap. I'd assumed opaque credit systems hid reasonable economics. Turns out they don't.
The Telegram angle is interesting though - moving agents into chat shifts who can actually use them. That's a bigger deal long-term than the quality gap between platforms.
How stable is Manus for recurring tasks after a few weeks? Is it drifting?
yes Jenny, I subscribed for a month only yo test out the Manus with Claude Code system and well, Wyndo is working out the comparison story on Perplexity Computer with it.. We both started with a vision of Perplexity vs Manus comparison only but this ended up being a dedicated article deserving its own place and we got the opportunity to have the series in place.. Soon Wyndo is going to release the other parts😀
You're all doing the lord's work testing, summarizing, and writing all these, Dheeraj and Wyndo. I've been curious about Manus for the longest time, but can't really find the time or motivation to try out yet another tool. This satisfied a lot of my curiosity.
that’s really nice to know James.. well yes it has been made easy to create agents but i still feel I am very happy with where Claude Code systems stand for me plus I felt it even though easy but costly too
What the heck, this is an amazing article. Dheeraj - the automation guru haha
Cant argue with this
ha ha, I will take that as super compliment not that it coming from two other gurus :) .. thanks guys
Thank you so much @Wyndo for giving me this opportunity to write about it.. It was so much fun researching, playing and writing about this comparison of Manus and Claude Code. I am still staying with Claude Code but it was so easy to use and setup AI Agents with Manus, it was pretty amazing..of course the the credits costs :)
awesome post!
Thank you for this test! For me, these are all use-cases where I use Perplexity. So basically all that needs web research in some form, I give to Perplexity (tasks, computer etc)
Especially some of my personal and health related things (this is why I'm looking forward to the new Health features in PPX with connectors to Oura etc). I recently started a new med for a chronic illness that also affects my sleep and I'm doing a daily diary with my Whoop and Oura data, my subjective experience and Computer is not only writing this in a diary, but also giving me the context based on pharmacology - basically being my doctor in my pocket.
But also for business stuff and research, I use PPX (with of course Claude as base model). So currently I haven't found a use-case where I would need OpenClaw or Manus.
Yeah i use perplexity computer too. Will have another deep dive for it in the upcoming post. But perplexity is kinda costly. For agentic task, it can spend hundreds of credits. But i kike that it can automate almost anything and have many app integration
Right, especially since you don’t get credits while being on the Pro plan - and the Max plan is totally unreasonable priced, especially when you got a Max plan already in Claude.
Exactly!
Sometimes hard to justify because u already have claude max plan and now even i’m expanding again to codex 😅
Oh, I just saw that Manus is part of Meta. Well, this would be an instant 'pass' for me anyway. 😅
yeah Manus is now part of Meta
Great use-case.
The 'control vs simplicity' split is real and I keep running into it. I use Claude Code for structured pipelines where I want to know exactly what the agent touched and why. The precision is worth the setup cost.
What surprised me about your breakdown: the 5-10x cost gap. I'd assumed opaque credit systems hid reasonable economics. Turns out they don't.
The Telegram angle is interesting though - moving agents into chat shifts who can actually use them. That's a bigger deal long-term than the quality gap between platforms.
How stable is Manus for recurring tasks after a few weeks? Is it drifting?
The gap between control and ease is where most teams get stuck.
Thanks, Dheeraj, really thoughtful breakdown.
I'm curious do you have a Manus subscription? Or how do you see the cost comparing to Perplexity?
Feels like that comparison would be interesting since both run on a credit system, but seem to behave very differently in practice.
yes Jenny, I subscribed for a month only yo test out the Manus with Claude Code system and well, Wyndo is working out the comparison story on Perplexity Computer with it.. We both started with a vision of Perplexity vs Manus comparison only but this ended up being a dedicated article deserving its own place and we got the opportunity to have the series in place.. Soon Wyndo is going to release the other parts😀
Looking forward to it!
You're all doing the lord's work testing, summarizing, and writing all these, Dheeraj and Wyndo. I've been curious about Manus for the longest time, but can't really find the time or motivation to try out yet another tool. This satisfied a lot of my curiosity.
that’s really nice to know James.. well yes it has been made easy to create agents but i still feel I am very happy with where Claude Code systems stand for me plus I felt it even though easy but costly too
Yeah, same. Very happy with Claude Code