Following OpenAI’s lead, Google released a slew of AI announcements ahead of the holiday season, focusing on agentic AI. The announcements included the highly anticipated Gemini 2.0, which is powering other advanced AI products, including Project Mariner.
Project Mariner is a research prototype built on Gemini 2.0 that explores human-agent interactions. It can automate tasks in your browser, acting like an AI agent by carrying out tasks delegated to it on the web for you. If it seems a bit complicated, the examples will really help bring it to life.
Also: Google’s Gemini 2.0 AI promises to be faster and smarter via agentic advances
How does it work? Project Mariner lives in the web browser as an experimental Chrome extension and understands the contents of your screen, including images, code, forms, and more. On the right-hand window, you can type into the chatbot whatever task you want the agent to perform for …