Manus' Wide Research: New Multi-Agent AI Tool Challenges OpenAI, Google
Manus, an AI startup that recently shifted its operations from China to Singapore, Tokyo, and San Mateo, California, has unveiled Wide Research, a new multi-agent AI tool designed to execute complex, large-scale tasks requiring extensive information gathering. The company announced Wide Research in a blog post on Thursday, July 31, marking it as Manus’s most significant launch since its debut in March of this year.
Wide Research is positioned by Manus as a sophisticated alternative to existing research-focused AI tools from its US counterparts, specifically citing OpenAI’s Deep Research and Google’s Deep Think. Its introduction also follows weeks after the debut of OpenAI’s ChatGPT Agent.
Manus stated that Wide Research “unlocks a powerful new way for users to tackle complex, large-scale tasks that require information on hundreds of items.” The company added that while Wide Research represents a milestone in its exploration of AI agent scalability, it is merely the beginning, with the underlying infrastructure being part of a larger ongoing development.
The new feature is immediately accessible to Manus’s paid Pro customers, with plans for a future rollout to Plus and Basic tier users.
Underlying Technology
Wide Research is the first feature from Manus built upon its extensive virtualization infrastructure and highly optimized agent architecture, a development process that reportedly spanned several months. At its core, Wide Research functions as a system for parallel processing and a protocol enabling collaboration among AI agents.
Unlike some other multi-agent systems, such as specialized AI agent managers or coding assistants, Wide Research distinguishes itself by employing general-purpose AI agents. This design choice grants the system significant flexibility, allowing tasks to transcend rigid formats or predefined domains. Manus suggests that the practical limitations are primarily confined to user imagination.
Practical Applications
In a demonstration video posted on X, Manus co-founder Peak Ji showcased Wide Research’s capabilities, illustrating how it can rapidly conduct research on 100 different sneakers simultaneously or generate 50 distinct poster designs within minutes.
Manus indicates that users will be able to undertake deep, high-volume research activities such as exploring Fortune 500 companies, comparing top MBA programs, or investigating various generative AI tools. However, Ji cautioned that Wide Research remains in an experimental phase and may present certain limitations.
Manus’s Expanding AI Portfolio
Earlier this year, Manus made an impact in the AI landscape with the introduction of its general-purpose AI agent. This agent is capable of autonomously performing web-based tasks, such as creating trip itineraries based on simple prompts. In June, the company further expanded its offerings with the release of its own AI video generator.
Manus’s AI features reportedly integrate with other large language models, including Anthropic’s Claude. The company explains that a dedicated cloud-based virtual machine runs behind every Manus session, enabling users to orchestrate complex cloud workloads simply by interacting with an agent. This underlying virtual machine provides Manus with its broad utility, facilitating diverse applications from generating customized rental presentations to securely evaluating open-source projects.
Manus currently does not offer its AI products and tools in China, according to a Bloomberg report. The AI startup recently secured $75 million in a funding round led by the US venture firm Benchmark. Its early investors include prominent Chinese tech entities such as Tencent, alongside HSG (formerly Sequoia China) and ZhenFund.