-
DOM fallback. When nothing is exposed for the job, the agent reads a text snapshot of the page and clicks/fills by element ref — the same trick page-agent popularized, no multimodal model needed.
-
Puter.js
Puter.js - The Backend for AI-Generated Apps. One-shot full-stack apps with your existing AI coding tool. Puter.js gives you Auth, Storage, DB, AI & more, with up to 90% fewer AI tokens than other backend platforms.
-
This is where gui-agent comes in, an open-source library I built exactly for this pattern (@aralroca/gui-agent on npm). It implements the emerging W3C WebMCP draft: your app registers its actions as tools on document.modelContext, and any WebMCP agent, yours, or eventually the browser's native one, can call them.
-
This is where gui-agent comes in, an open-source library I built exactly for this pattern (@aralroca/gui-agent on npm). It implements the emerging W3C WebMCP draft: your app registers its actions as tools on document.modelContext, and any WebMCP agent, yours, or eventually the browser's native one, can call them.
-
The brain runs in a small Node service built on Mastra. The agent itself is a few lines, model via OpenRouter, conversation memory in Postgres: