Manus, the viral AI “agent” platform from Chinese startup Butterfly Effect, has drawn attention to more than its own product. Its rise has also pushed Browser Use, a tool built for web agents, into a much brighter spotlight.
Browser Use is designed to help agentic applications work with websites on a user’s behalf. In practice, that means giving AI systems a clearer way to understand and operate web pages, from clicking menus to filling out forms.
A sharp jump in developer attention
The clearest signal is download growth. According to co-creator Gregor Zunic, daily downloads more than quintupled from around 5,000 on March 3 to 28,000 on March 10.
That change happened over a very short period, and Zunic described the moment as unusually intense. “The past few days have been really wild,” he said via DM. “We are the biggest trending repository [on GitHub], got loads of downloads [and] all that actually converts to big usage numbers.”
The surge appears closely tied to Manus. A post explaining how Manus uses Browser Use gained over 2.4 millions views and hundreds of reshares on X. That visibility helped move Browser Use from a developer tool with a specific purpose into a broader conversation about AI agents and web automation.
What Browser Use does for AI agents
Browser Use focuses on a practical problem: websites are built for people, while AI agents need structured ways to interpret and manipulate them. The tool extracts website elements such as buttons and widgets so AI models can interact with them more easily.
That matters because a web agent is only useful if it can reliably move through web apps and sites. Browser Use is built to support actions that look routine to a human user but can be difficult for software agents to handle consistently.
Its capabilities include:
- Managing multiple browser tabs.
- Handling mouse and keyboard inputs.
- Setting up actions such as saving files.
- Performing database operations.
- Helping agents click through site menus and fill out forms.
Manus uses Browser Use as one of its components for carrying out tasks. That connection gave developers a concrete example of how this kind of browser layer can fit inside a larger AI agent platform.
From ETH Zurich project to a company
Zunic launched the company behind Browser Use with Magnus Müller last year out of ETH Zurich’s Student Project House accelerator. Their starting belief was that web agents would be the “big thing” in 2025.
The early build moved quickly. “What started as casual brainstorming over a few lunches turned into a challenge: Let’s build something small, throw it on Hacker News, and see what happens,” Zunic said. “We put together an MVP in four days, launched it, and boom — number one. From there, it’s been an absolute rocket.”
That origin story helps explain the product’s position. Browser Use is not presented as a consumer-facing AI assistant. It is infrastructure for developers who want agents to operate across the web.
The company sells managed plans, but it also offers a free, self-hosted version of the software. The free, self-hosted version is the one that has blown up in the days since Manus was unveiled.
The web agent opportunity
Zunic has framed Browser Use as a tool for the builders chasing the web agent opportunity. He said he and Müller are trying to “sell a shovel” to developers moving into that space.
The ambition is to become a foundational layer rather than a single-purpose app. “We wanted to create a foundation layer that everyone will build browser agents on,” Zunic said. “In our minds, there will be more agents on the web than humans by the end of the year.”
That is a very strong forecast from the company’s co-creator, but the broader AI agent market is also expected to expand. Research and Markets says the sector will reach $42 billion in 2029. Deloitte anticipates that half of companies using AI will deploy AI agents by 2027.
Those projections help explain why a tool like Browser Use can gain traction quickly. If more companies and developers build agents that need to use websites, then the browser interaction layer becomes a key part of the stack.
Why the Manus effect matters
The Browser Use surge shows how attention can move through the AI ecosystem. Manus became the visible platform, but one of the tools underneath it also benefited when developers began examining how it works.
That kind of secondary momentum is important. AI agent products often depend on many enabling pieces: models, browser control, task execution, interface handling, and developer infrastructure. When one agent platform becomes widely discussed, the supporting tools can become more valuable to builders trying to create similar systems.
For Browser Use, the timing appears to have been fortunate. The company had already built a tool for web agents, offered a free self-hosted option, and positioned itself around a category that many expect to grow. Manus then gave the market a high-profile example of why such a tool might matter.
The result is a rapid jump in downloads, a top-trending GitHub repository, and a clearer place for Browser Use in the developing AI agent landscape.