A new Chinese artificial intelligence platform is causing a stir. But is it worth it? Techoreon takes a look.
A new Chinese Artificial Intelligence (AI) platform called Manus AI claims to have developed the world’s first fully autonomous AI agent, which is making waves around the world as an impressive new AI benchmark following the success of fellow Chinese company’s AI model, DeepSeek.
Its launch was announced on March 5 on the social media platform X by the startup Butterfly Effect, creator of Manus. The company described it as “the first general AI agent” that autonomously performs complex tasks. But it has already drawn criticism from experts who warn about data privacy. Here’s everything we know about Manus AI and what it can do.
What can Manus AI do?
Tech companies are scrambling to develop AI agents, which don’t have a specific definition but can generally be explained as having a degree of autonomy, meaning they can perform tasks like shopping, analyzing data, and generating reports based on instructions from a human rather than simply providing information.
Manus AI explains on its website that it takes its name from the Latin word for hand and that it’s “a general AI agent that turns your thoughts into actions.” The platform claims on its website that it can supposedly perform tasks such as purchasing properties, programming video games, analyzing stocks, and planning travel itineraries.
Manus’ creator claims in a video that it’s more than “just another chatbot or workflow… It’s a fully autonomous agent. We see it as the next paradigm of human-machine collaboration and potentially a glimpse of AGI,” said Manus AI chief scientist Yichao ‘Peak’ Ji, referring to artificial general intelligence, a type of AI that matches or exceeds human cognitive abilities and over which we could lose control.
The platform isn’t entirely developed from scratch, as it relies on existing large language models (LLMs) such as Anthropic’s Claude and Alibaba’s Qwen. However, Manus AI has garnered attention for its strong performance in benchmark tests, which show it outperforms OpenAI’s o3-based Deep Research agent, according to a chart posted on the official Manus website.
What do AI experts think?
Some AI experts have called Manus AI “China’s second DeepSeek moment,” according to Rowan Cheung, founder of The Rundown AI newsletter, in a LinkedIn post, who later tested the platform. “This AI agent called Manus is going crazy viral in China right now… It’s like Deep Research, plus Operator, plus Claude Computer combined, and it’s REALLY good,” he added.
But other experts say that, despite being heralded as another major AI breakthrough, Manus is unlikely to live up to the hype. “From what we’ve seen so far in Manus demonstrations, there’s a lot of evidence that Manus doesn’t get things right the first time or gets stuck in infinite loops,” Bradford Levy, an assistant professor at the University of Chicago’s Booth School of Business, told Euronews.
“Overall, there’s a notable lack of transparency about what Manus is actually capable of, suggesting that despite all the hype about Manus, there’s a good chance it won’t live up to expectations,” he added. Levy also warned that users should be wary of the data they provide to platforms like Manus and question its ties to the Chinese Communist Party.
China’s National Intelligence Law stipulates that companies must “support, assist, and cooperate” with state intelligence agencies, meaning any data shared on mobile and web apps can be accessed by those agencies. “By now, it’s well established that Beijing is actively working to covertly collect sensitive information on people around the world. What better way to achieve that end than to offer a brilliant app where people voluntarily hand over sensitive information?” Levy said.
Other Manus AI users said it displayed error messages and made mistakes and omitted information that could easily be found online. “Unfortunately, Manus AI crashed after 50 minutes at step 18/20. It was working quite well. It was working quite well – I looked at the Manus results and they seemed excellent. However, running the same question a second time is a bit frustrating as it takes too long,” said Pierre-Carl Langlais, co-founder of AI company Pleias, in a post on X.
After experimenting with it, I do like the UI. But it's fundamentally a workflow like Devin, not an actual agent (at least nothing really beyond the built-in agentic capacities of Claude)
— Alexander Doria (@Dorialexander) March 9, 2025How to Try Manus?
Trying Manus AI is unfortunately currently limited to an invitation-only phase due to its restricted server capacity. To get a chance to try it, request for invitation on the Manus AI website. This way, you’ll be among the first to know when new invitations are released and the platform moves toward broader public access.