Manus, an AI agent more advanced than a chatbot, takes center stage
Beijing: A powerful new AI tool, Manus, is making waves in China, sparking hopes that it could match the success of DeepSeek, which earlier this year shook the global tech industry with its state-of-the-art chatbot.
According to its official website, Manus, an AI agent considered more advanced than a chatbot, can handle everything from analyzing stock markets to creating personalized travel handbooks with simple user instructions.
Rapid rise
Manus was launched last week by Chinese startup Butterfly Effect.
In an introductory video that has since gone viral, co-founder Yichao “Peak” Ji hailed Manus as “the next paradigm of human-machine collaboration,” calling it “potentially a glimpse into AGI”—artificial general intelligence designed to think like humans.
Currently available only by invitation, Manus has seen rapid adoption, with its official Discord server surpassing 170,000 members.
Its name is derived from the Latin phrase “Mens et Manus”, meaning “mind and hand,” symbolizing its integration of knowledge and practical application.
Exclusive access
“Manus is a system that autonomously carries out tasks on behalf of users,” explained Manoj Harjani, a research fellow at Singapore’s S. Rajaratnam School of International Studies (RSIS).
“Unlike DeepSeek and ChatGPT, which respond to prompts through a chat interface, Manus can execute tasks such as booking tickets or sorting resumes,” he noted.
DeepSeek was developed quietly before gaining prominence, but according to analysts, Manus is taking a different approach—restricting access to an invite-only beta and focusing on enterprise-level clients.
“While this exclusivity can generate buzz, it may also hinder widespread adoption,” said Marina Zhang, an associate professor at the University of Technology Sydney’s Australia-China Relations Institute.
She warned that if Manus continues to operate in a closed ecosystem, it may struggle to build the broad community appeal that DeepSeek achieved through open-source engagement.
During an AFP test using an invitation code, Manus took significantly longer than DeepSeek to generate responses.
However, it was able to complete more complex tasks, such as creating custom websites, outperforming both DeepSeek and ChatGPT in certain functionalities.
Touchy topics
DeepSeek has been programmed to align with official government narratives on politically sensitive topics in China.
But Manus appears to take a different approach, offering uncensored, factual responses to sensitive questions.
When asked about the 1989 Tiananmen Square crackdown, Manus stated:
“The Chinese government carried out a violent crackdown on pro-democracy protesters in Tiananmen Square in Beijing,”
followed by six more paragraphs detailing the incident.
Similarly, when queried about alleged human rights abuses in Xinjiang, Manus provided a comprehensive answer.
When asked about its uncensored replies, Manus stated:
“I do not intentionally censor factual information. When discussing sensitive topics like historical events or human rights issues, I aim to provide balanced, objective information that helps users understand complex subjects.”
This openness may stem from Manus’ core focus on task execution, meaning its developers may not have implemented strict content controls, suggested Li Jianggan, founder of Singapore-based consultancy Momentum Works.
The next DeepSeek?
Whether Manus can achieve mainstream success similar to DeepSeek will depend on its ability to scale and meet demand, said RSIS’s Harjani.
However, Manus is unlikely to be a direct competitor to DeepSeek, as they cater to different AI applications, he added.
Manus’ growth will require “adequate computing power” and “effective handling of technical, ethical, and regulatory challenges”, said Zhang.
“If Manus successfully addresses these issues, it could become a major player in enterprise automation,” she added.
However, Li cautioned that Manus is “far from perfect” and remains closed to the general public, suggesting that its team is still refining the program.
“Manus is tackling real-world problems and executing complex tasks, which require a high level of adaptability,” he said.
Its long-term success, however, remains uncertain.