At the 2025 Google I/O Developer Conference, Google made a significant announcement: the expansion of its experimental AI agent project, Project Mariner, to a broader audience of users and developers. This project, tailored for browsing and interacting with websites, has undergone substantial enhancements and now boasts the capability to monitor and perform up to 10 tasks concurrently. Once users demonstrate a specific task to Mariner, it learns and automates it for future use. The Mariner tool for computers is currently accessible via the Gemini API, with the agent mode poised to debut soon on Chrome, Search, and Gemini applications.
