Google shuts down Project Mariner as AI agent race takes a new direction

Google shuts down Project Mariner

Google has pulled the plug on Project Mariner, an experimental feature designed to perform tasks for you across the web, as reported earlier by Wired's Maxwell Zeff. The Project Mariner landing page now contains a message that says: "Thank you for using Project Mariner. It was shut down on May 4th, 2026 and its technology voyaged to other Google products." Google first revealed Project Mariner in December 2024 and later announced an update allowing it to perform up to 10 tasks at a time. Over the past year, Google has integrated features powered by Project Mariner into its other AI tools, including Gemini Agent, which can do things like archive emails on your behalf or help you book a hotel. Google also folded Project Mariner's agentic capabilities into its AI-powered search feature, AI Mode.

Google Closes Project Mariner, Web Browsing AI Shut Down Earlier This Week

(Credit: Cole Kan | Image Credit: Google/ Chorna Olena, Daniel Megias,da-kuk via Getty Images) One of Google's first forays into agentic AI for web browsing has been closed without any prior warning. Project Mariner closed on Monday, May 4, according to Google's website, as spotted by Wired's Max Zeff. Google's landing page for the experimental tool now says, "Thank you for using Project Mariner. It was shut down on May 4th, 2026 and its technology voyaged to other Google products." It tells users to try using Gemini Agent for what it calls "complex tasks." Project Mariner tools enabled web browsing via an AI agent you could send to complete tasks for you across the internet, taking the effort of searching out of your hands. The idea is to make deep research tasks easier, such as finding the best-priced flights, planning a hotel stay, or researching a new purchase. The tool launched in December 2024 before becoming a greater focus during last year's Google I/O. Project Mariner expanded to handle 10 tasks simultaneously, though you needed a $249.99/month Google AI Ultra subscription to access it. Other Google AI tools, such as Gemini Agent and search tools baked into AI Mode, provide similar functionality to Project Mariner, but are much easier to access. Google's "auto-browse" feature in Chrome, which first debuted early this year, can also handle similar, complex web browsing tasks. Google has chosen to shut down Project Mariner within a few weeks of Google I/O 2026, set for May 19. That may be a sign Google is clearing out its experimental tools ahead of the next event. However, it's still unlike Google to shut down a tool without a prior announcement. Zeff previously reported that Google moved Project Mariner staff to a new team working on an upcoming OpenClaw-like agent. If work on the tool has progressed, Google may be preparing to announce its next agentic tool soon to help compete with other AI rivals.

Project Mariner is dead, but Google's browser-controlling AI plans are not

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. What just happened? Google has quietly killed off Project Mariner, one of the first AIs designed to automatically control web browsers to retrieve information, make purchases, and perform other actions - a browser-based AI agent, essentially. But this is unlikely to be considered a failure by the company as many of Mariner's features have been folded into other Google products since its launch. Google first announced Project Mariner back in December 2024. An extension for an experimental build of Chrome, Mariner could execute multi-step commands to browse websites, use Google search, retrieve specified information, go shopping, and more. Google positioned the agent as assisting with tasks that are usually tedious for humans. Mariner was later updated to handle 10 tasks simultaneously, but it could only be accessed via a $250-per-month Google AI Ultra subscription. Mariner worked by taking screenshots, identifying text and buttons, then clicking and typing like a human, a process that can be slow, expensive, and prone to mistakes. Google's landing page for the experimental AI tool confirms it was shut down on May 4, 2026, and its technology moved to other Google products. One of these was Gemini Agent, which the Mariner page suggests using for "complex tasks." Wired's Max Zeff, who spotted the notice, reported in March that Google was moving staff off Project Mariner and onto a new web-browsing AI agent designed to compete with similar tools from OpenAI, OpenClaw, and Perplexity. Adoption of consumer browser agents has been underwhelming, especially compared with the rapid rise of coding-focused tools such as Claude Code and more flexible autonomous agents such as OpenClaw. OpenClaw's success also underlines the problem for browser agents: people like the idea of an assistant that can shop, book travel, browse the web, but in practice they often want something faster, safer, and more flexible than a bot cautiously clicking around websites. That helps explain why Mariner's shutdown feels more like a pivot than a surrender. The concept is still alive inside Gemini Agent and AI Mode, just without the standalone branding or the $250-per-month Ultra-only experiment attached to it.

Google quietly kills Project Mariner as the AI agent race shifts gears

Google isn't abandoning the tech entirely, as many of the tool's features are being folded into the Gemini API and Gemini Agent. Google is reportedly pulling the plug on Project Mariner, the experimental AI browser agent it once positioned as the future of how people interact with the web. If you didn't follow the buzz, Project Mariner (introduced during I/O 2025) was Google DeepMind's effort to create an AI that could use the internet much like a person. However, in March, there were signs that the project was going to bite the dust after Google moved staffers off the Project Mariner team, as reported by Wired's Maxwell Zeff. X user BoughtMilkMan first spotted the project's shutdown, as indicated by its landing page's message. Unlike basic chatbots that just summarize web pages, Mariner could move through Chrome, fill out forms, look up job listings, and even book travel on sites such as Expedia. It did this by taking frequent screenshots of your browser, recognizing buttons and text, and then clicking or typing for you. Although the concept of an AI assistant browsing the web is appealing, the technology is resource-intensive. These agents require significant computing power to process visual data in real time, which can result in slow performance and occasional errors, such as selecting incorrect options. There's another reason for this change: the industry has moved forward. While Google worked on browser-based agents, new agentic AI tools like OpenClaw and Claude Code have become more popular. These tools go beyond clicking links -- they can modify files, write complex code, and act as digital coworkers. If you liked what Mariner could do, there's no need to worry. Google has said that Mariner's technology will live on in other products. Most of its features are being added to the Gemini API and the new Gemini Agent.

Google pulls the plug on Project Mariner, the AI agent that browsed the web like a human

The autonomous browser agent Google introduced at I/O 2025 is now more. Its technology is being absorbed into the Gemini API and Gemini Agent. Google has shut down Project Mariner, the autonomous web browsing agent it debuted at I/O last year. The tool, which could navigate Chrome, fill out forms, search listings, and book travel by taking screenshots and visually recognizing page elements, is no longer available. Its landing page now shows a notice with the shutdown date listed as May 4, 2026. A browser agent that saw what you saw Project Mariner was Google DeepMind's attempt to build an AI agent that interacted with websites the way a person would. Rather than reading page data directly, it processed screenshots in real time to identify buttons, text fields, and links, then clicked and typed on a user's behalf. That approach lets it handle multi-step tasks across sites without requiring any special integration from the website. The tradeoff was performance. Visual processing at that scale demands significant compute, and the method was prone to errors, such as selecting the wrong option on a page. The agent also raised privacy concerns, since it required continuous access to whatever was visible in a user's browser at any given moment. Recommended Videos Signs of trouble first surfaced in March, when Wired reported Google had begun reassigning staffers away from the Project Mariner team, a signal the project was losing internal support months before the shutdown became public. Mariner's tech isn't going away Google says Mariner's technology "voyaged to other Google products." Its core features will reportedly be absorbed into the Gemini API and the new Gemini Agent rather than being discontinued outright. The shutdown tracks with a wider shift in how the industry is building agentic AI. Tools that operate at the file and code level, rather than the visual browser level, have become the dominant model. They are faster, cheaper to run, and more capable of handling complex, multi-step tasks. Mariner's screenshot-based approach, while novel at launch, was competing against an architecture that had effectively moved past it.