AI Browser Agent: Gemini 2.5

Execute autonomous web tasks using Google's Gemini 2.5 computer use model

AI Browser Agent: Gemini 2.5 demo
TypeScript
Source code
npx create-browser-app --template gemini-cua
uvx create-browser-app --template gemini-cua

Run autonomous browser tasks with Gemini 2.5

Execute web tasks autonomously using Google's Gemini 2.5 computer use agent. Give the agent a natural language instruction and it navigates pages, performs searches, clicks through flows, and returns structured results, all without manual scripting. Built on Browserbase for reliable cloud browser execution.

Steps

  1. Uses Stagehand Agent with the Gemini 2.5 Computer Use to control the browser
  2. Navigates the web and performs the given instruction autonomously
  3. Returns the final result after completing all steps