alibaba/page-agent
Page Agent
Page-agent is an LLM browser automation agent and JavaScript in-page GUI controller. It translates natural language instructions into direct browser interface actions to automate web-based tasks and manipulate web page elements through a programmable interface.
The system coordinates complex sequences of actions across multiple browser tabs and different websites. It functions as a remote browser control server, providing an interface that allows external clients to operate a browser and manage page interactions.
Its capabilities include natural language intent decoding and action mapping, DOM-tree accessibility parsing, and cross-tab state management. The project also incorporates a JavaScript injection runtime and a remote procedure call interface to facilitate external command execution and state updates.
Features
- Autonomous Browser Agents - An intelligent agent that interprets natural language to navigate and interact with web interfaces.
- Browser Automation Agents - An LLM-powered agent that translates natural language into direct browser interface actions.
- Intent-to-UI Action Mappings - Maps natural language instructions to a sequence of executable browser operations and interface elements.
- Natural Language Command Translation - Translates natural language instructions into executable browser-level commands via LLMs.
- Natural Language Query Interfaces - Processes user text inputs to determine specific goals and parameters for web interface interaction.
- Natural Language Workflow Builders - Converts plain text instructions into executable sequences of browser agent actions.
- Accessibility Tree Extractions - Extracts the structural and semantic hierarchy of web page elements to identify interactable targets.
- Natural Language Automation - Controls web interfaces using text commands to automate repetitive browser workflows.
- In-Page GUI Controllers - Provides a JavaScript-based system for manipulating web page elements through a programmable interface.
- JavaScript Injections - Injects JavaScript directly into the browser context to manipulate the DOM in real time.
- Remote Browser Controllers - Provides a server interface for external agents to remotely operate browsers via API calls.
- Multi-Model Workflow Coordinators - Coordinates automated sequences that complete multi-step processes across various web interfaces.
- Multi-Page Browser Workflow Automators - Coordinates complex sequences of actions across multiple browser tabs and different websites.
- Remote Procedure Call Interfaces - Exposes server endpoints for external clients to send commands and receive browser state updates.
- Cross-Page Workflow Orchestration - Coordinates complex sequences of browser actions across multiple tabs and different websites.
- Cross-Tab State Coordination - Coordinates interaction sequences across multiple browser windows to execute multi-domain workflows.
