New AI Tools: Trae-Agent, Windows-MCP, Chrome MCP
AI automation tools: Windows-MCP controls PC operations, Chrome MCP automates browser tasks, and Trae-Agent handles software engineering workflows.
"AI Disruption" Publication 7100 Subscriptions 20% Discount Offer Link.
Gemini CLI - You can use it as a programming assistant, handle automation tasks, generate images and videos with MCP, and more.
With Windows-MCP added, you can also have it directly operate your Windows system.
A user's video shows a prompt like this:
"Using only Windows MCP, open Chrome, browse x.com, open my latest followers, and tell me the @ of the last person who followed me."
You can also remotely control your computer through SSH and have it do work.
Gemini: "Open Firefox and go to YouTube Music."
By streaming Gemini CLI to a browser, you can remotely control any device from any other browser-supported device.
In the video, a smartwatch was used to control a smart TV... without any setup needed on the watch.
The method used to stream Gemini CLI can allow anything to control the Gemini CLI on your computer, thereby controlling the computer (install various MCPs previously introduced to Gemini CLI to unlock more possibilities):
Basically it works like this:
Install Gemini CLI on Machine A
Expose your terminal to the web (Warning: risky)
Open a browser on Device B and go to Machine A's IP - control Machine A from Device B
Tool used: ttyd - Share your terminal over the web: https://github.com/tsl0922/ttyd
Gemini CLI's Gemini 2.5 Pro model tool calling might not be perfect - not as accurate as Claude's calling, so you can also use Claude Desktop or Claude Code + Windows-MCP, though the latter has a higher barrier to entry.