Google has officially announced the shutdown of its experimental web automation project, Project Mariner. Although the project is no longer operational, the technology it used has not been discarded but has instead been integrated into Google's other core products, such as Gemini Agent and AI Mode.
Project Mariner first appeared in December 2024 and was mainly used to perform multi-step tasks across websites, including information retrieval and operations. After multiple iterations, the project has developed the ability to handle ten tasks simultaneously, demonstrating strong multi-tasking capabilities. Over the past year, Google has gradually integrated the core functions of Project Mariner into its own product ecosystem to enhance its intelligence level.
In particular, its agent capabilities have been integrated into Gemini Agent, enabling it to handle tasks such as archiving emails and booking hotels. Additionally, these features have been introduced into Google's search product AI Mode to enhance the search experience.
In this context, Google recently launched a new AI feature in Chrome browser called "auto-browse." This feature can perform multi-step tasks, such as checking flight prices, aiming to directly compete with web proxy tools from competitors like OpenAI and Perplexity.
Key Points:
🌐 Google announced the shutdown of Project Mariner, but the technology has been integrated into Gemini Agent and AI Mode.
🤖 Project Mariner has the ability to handle ten tasks at the same time, enhancing multi-tasking capabilities.
✈️ The newly launched "auto-browse" feature can perform multi-step tasks, such as checking flight prices.
