This new Google Gemini model scrolls the internet just like you do - how it works

▤ Summary

Google has introduced Gemini 2.5 Computer Use, a model that can autonomously browse websites and perform user-directed actions step by step. The model works through an iterative loop that tracks recent UI actions, fetches pages and screenshots, and decides on the next action based on the interface state. Google says the preview is available through the Gemini API in Google AI and Vertex AI, with a demo also accessible via Browserbase. The release includes safety controls to stop risky actions and require confirmation for sensitive tasks. The model enters a competitive market already populated by similar web-browsing tools from OpenAI and Anthropic.

▦ Linked Companies

Browserbase

$1M to $5M