China is again establishing itself as a powerhouse in artificial intelligence with Manus, an advanced autonomous AI developed by startup Monica. Following the recent success of DeepSeek, Manus promises to compete directly with global leaders such as ChatGPT and Google, bringing AI autonomy to levels close to what is considered a level 3 autonomous agent, capable of independent planning, action, and decision-making.
Monica: The startup behind Manus
Founded in July 2023 by Xiao Hong, Monica started as a browser extension that integrated advanced generative AI models, such as Claude 3.7 Sonnet, Ask OpenAI o1, DeepSeek-R1, o3-mini, Veo 2 and Pix 2, simplifying everyday tasks for millions of users. It quickly attracted the attention of major Chinese technology investors such as Tencent, consolidating its position in the market. The initial success propelled Monica towards an even more ambitious goal, launching Manus in March 2025.
Who is behind Manus?
Xiao Hong: the pragmatic visionary
Xiao Hong, a young serial entrepreneur and founder of Monica, is known for developing successful technology solutions on platforms such as WeChat, attracting investment from giants such as Tencent and ZhenFund. His trajectory led him to create Butterfly Effect in 2022, where he launched Monica, always showing a pragmatic vision focused on valuable and practical solutions.
Ji Yichao: the technological brain
Ji Yichao, co-founder and chief technology officer, provides the technical depth that enables Manus to operate as an advanced autonomous agent. Recognized for previous projects such as Magi, an extensive knowledge graph that endows Manus with deep, autonomous reasoning capabilities, Ji is key to the technical execution of the project.
What makes Manus special?
Manus represents a significant advance towards a level 3 autonomous agent capable of performing complex tasks without constant human supervision. What sets Manus apart is its actual ability to plan, execute, and deliver complete results through a multi-agent structure that simulates human collaboration:
- Planning Agent: Divides tasks into manageable subtasks and creates execution strategies.
- Executing Agent: Performs practical actions such as advanced searches, financial analysis or creation of complete reports.
- Verification Agent: Supervises and validates results to ensure their quality.
Manus also has a powerful continuous learning capability, adapting to the user’s individual preferences.
Real-world examples of Manus use
https://manus.im/Laversatility and autonomy of Manus have been demonstrated in several real applications:
- Real estate search: Analyzes prices, security, and nearby schools and generates complete reports with personalized recommendations.
- Travel planning: Complete itineraries, reservations, and customized traveler quotes.
- Advanced financial analysis: Generates comprehensive reports on economic trends, interactive graphs, and strategic conclusions.
- Creation of educational materials: Develops interactive and dynamic content to explain complex secondary and higher education concepts.
Is Manus really standalone or just a smart integration?
The arrival of Manus has generated excitement, but also debates about whether it represents true autonomous AI or simply an efficient combination of existing technologies. Although Manus has demonstrated impressive capabilities in benchmarks such as GAIA, some experts argue that it has yet to fully prove its autonomy in real, complex contexts.
In addition, there are significant security and privacy concerns, given the extensive access Manus has to sensitive data and its ability to install itself on devices and control browsers and other computer systems. This raises a priori critical questions:
- What guarantees does Monica offer about the protection and privacy of its users’ data?
- How do you prevent abuse, cyber attacks or malicious use of the tool?
- What controls are in place to ensure that this powerful AI is not used for improper purposes?
Security and Privacy: Manus' great challenge
Manus’ advanced autonomy, while promising, poses significant privacy and IT security risks. Companies and users must question how Monica will manage the information collected and whether it will be transparent about using and storing personal and business data. In addition, the fact that Manus can act directly on external systems requires strict security measures to prevent abuse or cyber-attacks.
Manus represents an exciting new stage in the race for autonomous artificial intelligence. However, its true success will depend on how Monica addresses critical security, privacy, and scalability challenges and whether it transparently demonstrates true autonomy.
While innovation is evident, Manus has yet to prove that it is more than a sophisticated integration of existing technologies. Competition with Western tech giants will remain fierce, and the market will ultimately decide whether Manus is the promised technological revolution or just another interesting step in the ongoing evolution of artificial intelligence.
Manus AI is currently in a limited-access phase. It is available by invitation only. You can request it from their website.
Warning: there have been reported cases of resale of Manus AI access codes for high amounts on unofficial platforms.
Have a good week!
