Repolode
Submit
alibaba

alibaba/page-agent

A JavaScript library that embeds a GUI agent in any web page, letting users control the interface with natural language.

View on GitHubWebsite
Stars
29,513
Forks
2,879
Open issues
112
Last commit
9 hours ago

What is it?

Page Agent is a JavaScript library that runs inside a web page and gives it an AI agent that can operate the page's interface from natural language instructions. It needs no browser extension, Python runtime, or headless browser, since everything runs as in-page JavaScript.

The agent works on a text-based representation of the DOM rather than screenshots, so it does not require multi-modal LLMs. It can be connected to most mainstream models, including locally deployed ones, and is installed either via a script tag or the npm package.

For tasks spanning multiple pages, an optional Chrome extension is available, along with an MCP server (in beta) that lets external agent clients control the browser.

Key features

  • 01Runs entirely in page JavaScript, with no browser extension or backend required for basic use.
  • 02Operates on a text-based DOM representation without screenshots or multi-modal models.
  • 03Works with mainstream LLMs, including locally deployed models.
  • 04Optional Chrome extension supports multi-page tasks.
  • 05Beta MCP server lets external agent clients control the browser.

Use cases

  • Adding an AI copilot to a SaaS product without rewriting the backend.
  • Turning multi-step form workflows in ERP, CRM, or admin systems into a single sentence.
  • Making a web application operable through natural language or voice input.

Quick start

  1. 1

    Try instantly with the demo script tag (evaluation only)

    <script src="https://cdn.jsdelivr.net/npm/page-agent@1.12.4/dist/iife/page-agent.demo.js" crossorigin="anonymous"></script>
  2. 2

    Install the npm package

    npm install page-agent
  3. 3

    Import PageAgent and run it with your own LLM API key

Things to know

  • The demo CDN uses a free testing LLM API and is for technical evaluation only.
  • Using it with your own models requires an LLM API key or a locally deployed model.
  • It works in-page only, so multi-page tasks need the separate Chrome extension.

Topics

#ai-agents#browser-automation#javascript#typescript#llm#mcp#web#agent#ai

FAQ

What is page-agent?
A JavaScript library that embeds a GUI agent in any web page, letting users control the interface with natural language. Page Agent is a JavaScript library that runs inside a web page and gives it an AI agent that can operate the page's interface from natural language instructions. It needs no browser extension, Python runtime, or headless browser, since everything runs as in-page JavaScript.
Is page-agent free to use? What license does it have?
page-agent is open source under the MIT license. Check the license terms before using it in commercial products.
Is page-agent still maintained?
Actively maintained. The last commit to page-agent was on Oct 19, 2026.
What can I use page-agent for?
Adding an AI copilot to a SaaS product without rewriting the backend. Turning multi-step form workflows in ERP, CRM, or admin systems into a single sentence. Making a web application operable through natural language or voice input.
How do I install page-agent?
Try instantly with the demo script tag (evaluation only): <script src="https://cdn.jsdelivr.net/npm/page-agent@1.12.4/dist/iife/page-agent.demo.js" crossorigin="anonymous"></script> Install the npm package: npm install page-agent Import PageAgent and run it with your own LLM API key
Where can I get page-agent?
The source code and installation instructions are on GitHub: https://github.com/alibaba/page-agent

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects