Gemini vs Yad

Gemini has no write-up on Specifiction yet; its figures are above.

Yad turns plain-language instructions into actions in your own logged-in browser. It can click, type, navigate websites, fill forms, and handle research tasks, with human approval steps and scheduled tasks. The extension works in Chrome and Brave alongside a free companion app that currently runs only on Windows. You can bring an AI provider key or use a local model; listed options include Groq, Google Gemini, OpenRouter, and custom endpoints. Yad stores selectors, past runs, and learned fixes on your device for reuse, and the companion exposes a local API accepting goals at POST localhost:3747/goal. The maker says passwords, card numbers, and sensitive page fields are masked before content reaches an AI; local-model use keeps page content on the computer. The companion encrypts the AI key with Windows DPAPI, while the extension stores only the encrypted blob. Spending caps and one-click stopping are available, and payment or checkout pages are blocked from autonomous action. Hosted team setups require contacting Yad and cannot use personal logged-in sessions.