General Purpose
Open Source
Enables natural language commands to perform tasks across Windows applications.
About
UFO is a UI-Focused multi-agent framework designed to execute user requests on Windows OS by seamlessly navigating and operating within individual or multiple applications. It employs a dual-agent system to observe and analyze graphical user interfaces (GUIs) and control information of Windows applications, enabling the agent to perform tasks based on natural language commands.
Features
- Translates natural language commands into actionable operations on Windows OS.
- Enhanced by Retrieval Augmented Generation (RAG) from various sources, including offline documents and online search engines.
- Supports comprehensive automation with a diverse set of skills such as mouse, keyboard, native API, and "Copilot".
Tags
Windows Automation
UI Interaction
Multi-Agent Framework