Introducing Manus: The World's First Universal AI Agent - Another Chinese AI Product Splashes

Manus, claimed as the world's first universal AI Agent, officially launched a partial internal beta test, marking a crucial step forward in the field of AI autonomous intelligence. It is another Chinese AI product that has made a big splash, following DeepSeek.

According to the introduction, manus is a truly autonomous AI agent. With its powerful capabilities of independent thinking, planning, and executing complex tasks, it can directly deliver complete results, demonstrating unprecedented versatility and execution ability. Just as the name Manus implies, it symbolizes 'hand' in Latin. That is to say, knowledge should not only stay in the mind but also be executable by 'hand'. This is the essential advancement of Agent compared to AI Bot (chatbot) products.

Many users commented on social media that this is yet another sleepless night in the tech circle since DeepSeek made its splash.

To showcase the extensive application capabilities of Manus, the official team has specifically released 40 cases, covering various fields such as personalized travel planning, stock analysis, educational course development, insurance policy comparison, B2B supplier procurement, financial report analysis, company list organization, online store operation analysis, making activity explanatory diagrams, arranging candidate interviews, finding potential customers, and creating teleprompters for press conferences. These cases fully demonstrate the high execution efficiency and practicality of Manus in different scenarios.

In the authoritative gaia benchmark test, Manus has set a new record, far surpassing similar products of OpenAI in performance, further consolidating its leading position in the field of AI autonomous intelligence.

To sum it up in one sentence - what Manus really wants to be is your literal 'agent' in the digital world. And it has achieved that.

'Digital Agent'

Firstly, the biggest difference between Manus and previous Large Language Models (LLMs) in terms of user experience is that it emphasizes the ability to directly deliver the final result, rather than just providing a simple 'answer'.

Manus currently adopts a Multiple Agent architecture and operates in a similar way to the Computer Use released by Anthropic previously, running completely in an independent virtual machine. At the same time, it can call various tools in the virtual environment - writing and executing code, browsing web pages, operating applications, etc., and directly delivering complete results.

In the official released video, three work cases completed by Manus in actual usage scenarios are introduced:

The first task is resume screening.

From 15 resumes, it recommends suitable candidates for the position of reinforcement learning algorithm engineer and ranks the candidates according to their reinforcement learning expertise.

In this demonstration, you don't even need to unzip the compressed file or manually upload each resume file one by one. At this time, Manus has already shown its human-like 'intern' side. It manually unzips the file and browses each resume page by page, while recording the important information in it.

In the result provided by Manus, there is not only an automatically generated ranking suggestion, but it will also divide the candidates into different levels according to important dimensions such as work experience. After receiving the user's preference for presenting the result in the form of an Excel table, Manus can automatically generate the corresponding table by writing a Python script on the spot.

Manus can even use its memory ability to record information like the user prefers to receive the result in the form of a table during this practical process. The next time it processes similar task results, it will preferentially present them in the form of a table.

The second case is property selection.

In the case, the user wants to buy a property in New York and the input requirements are a safe community environment, a low crime rate, and high-quality primary and secondary school education resources - of course, including the most important budget, which should be affordable within the monthly fixed income.

For this requirement, the Manus AI breaks down the complex task into a to-do list, including researching safe communities, identifying high-quality schools, calculating the budget, searching for properties, etc. And through web searches, it carefully reads articles about the safest communities in New York and collects relevant information.

Secondly, Manus writes a Python program to calculate the affordable property budget according to the user's income. Combining with the relevant housing price information on real estate websites, it screens the property list according to the budget range.

Finally, Manus will integrate all the collected information and write a detailed report, including community safety analysis, school quality assessment, budget analysis, recommended property list, and relevant resource links - just like a professional real estate agent. And because Manus has the attribute of 'fully considering the user's interests', its usage experience is even better.

In the last case, Manus demonstrates its ability to analyze stock prices.

The task given in the case is to analyze the correlation between the stock prices of NVIDIA, Marvell Technology, and Taiwan Semiconductor Manufacturing Company (TSMC) over the past three years. As is well known, there is a close correlation among these three stocks, but for novice users, it is difficult to quickly sort out the causal relationship.

The operation of Manus is very similar to that of a real stockbroker. It first accesses information websites such as Yahoo Finance through the API to obtain historical stock data. At the same time, it will also cross-verify the accuracy of the data to avoid being misled by a single information source, which may have a significant impact on the final result.

In this case, Manus also uses its capabilities of writing Python code, conducting data analysis, and visualization. It also introduces professional financial tools for analysis. Finally, through data visualization charts and a detailed comprehensive analysis report, it feedbacks the causal relationship to the user - really just like the daily work done by an 'intern' in the financial field.

Moreover, the official website of Manus also showcases more than a dozen scenarios where Manus can be used: you can directly use Manus to organize your itinerary, get personalized travel route recommendations, and even let it learn to use various complex tools to complete daily work in a streamlined manner.

During this process, what really makes Manus stand out from previous tools is its autonomous planning to ensure the ability to execute tasks.

Behind Manus is Monica.im

In 2024, as soon as GPT-4o, Claude 3.5, and the OpenAI o1 series were launched, Monica enabled users to access the latest State-of-the-Art (SOTA) models. With the new progress in model access, the professional search, DIY Bot, Artifacts for writing mini-programs, memory, and other functions launched by Monica have also been loved by users. And Monica presents different interactive forms and functions on web pages of different functions such as YouTube, Twitter, Gmail, and The Information, to adapt to the user needs of specific scenarios, and has updated the personalized AI experience of hundreds of web pages.

In 2024, the number of Monica users doubled to 10 million. At the same time, it maintained considerable profitability and ranked among the top in similar overseas products.

The R&D team behind Manus is from Monica.im, and its founder, Xiao Hong, is a young serial entrepreneur. Born in 1992, he graduated from Huazhong University of Science and Technology. In 2015, after graduating, he started his own business, and his early entrepreneurship was not very smooth. In 2016, he started a business to provide editing and data analysis tools for WeChat official account operators, which attracted millions of users, achieved profitability, and the product was finally sold to a unicorn company in 2020.

After the wave of large language models in 2022, he officially founded Monica, focusing on the overseas market. Through the product ChatGPT for Google, an independent developer product, the product quickly completed its initial growth.

When interviewed by the media, Xiao Hong said that products should not only exist in the form of chatbots. Agent will be a new form, and new products are needed to undertake it.

He got inspiration from the AI programming products cursor and Devin. An Agent should also, like Devin, be oriented towards the general public and truly be executed under the guidance of AI. However, the problem in the past was that the models were not intelligent enough.

But perhaps encapsulating services for scenarios based on the existing capabilities of the models is exactly the advantage of the Monica team. Xiao Hong said that currently, there are not many Agent product teams because it requires a lot of comprehensive capabilities. For example, the team should have experience in chatbots, AI programming, and browser-related fields (because they all run on browsers), and have a good sense of the boundaries of the models - what level it has developed to today and what level it will develop to in the future, etc.

He believes that the Agent is still in its early stages. Firstly, the Agent is still in the planning stage and has not reached the stage of execution in the physical world. Secondly, the capabilities of large language models are still evolving, and everything is still unpredictable.