PageAgent.js is a front-end JavaScript library from Alibaba that embeds a GUI AI agent directly into any webpage with a single script tag. It requires no backend, no Python, and no headless browser. Users can issue natural language commands to control the UI, and the agent executes actions with a human-in-the-loop collaborative panel. It supports multiple LLM providers including OpenAI, Claude, DeepSeek, Qwen, and offline via Ollama. Use cases include SaaS AI copilots, smart form filling, and accessibility improvements. An optional browser extension enables multi-page and multi-tab task execution.

2m read timeFrom alibaba.github.io
Post cover image

Sort: