Page Agent
by Alibaba
Embeds a GUI agent in any web page with one script tag, driving the existing DOM from natural-language instructions without screenshots or a headless browser.
Skills
In-Page Task Execution
Executes natural-language instructions on the current page by reading a text-based DOM tree, then clicking, typing and scrolling.
Smart Form Filling
Collapses long click-through workflows in ERP, CRM and admin systems into a single sentence that fills and submits page forms.
Bring Your Own Model
Connects to mainstream or locally deployed models through an OpenAI-compatible baseURL and API key set at agent construction.
Multi-Page Tasks
Extends a task across browser tabs using the optional Chrome extension, coordinating steps that span more than one page.
MCP Remote Control
Exposes the in-page agent through a beta MCP server so external agent clients can drive the user own browser session.
Related Agents
Browser Harness
Connects an LLM directly to a real Chrome browser over a single CDP websocket, letting the agent write the helper code…
OpenHands
Software engineering agent platform — runs in sandboxed Docker containers, edits code, runs tests, fixes bugs, and open…
Safari MCP Server
Connects AI agents to a live Safari window for web debugging: DOM inspection, network monitoring, console logs, screens…
OpenSandbox
Runs AI-agent workloads in isolated Docker or Kubernetes sandboxes, exposing sandbox lifecycle, command, filesystem, an…