UFO²
UFO² is Microsoft's next-generation Desktop AgentOS that transforms natural-language requests into automatic, reliable, multi-application workflows on Windows, combining UI automation, native API integration, and multi-agent coordination.
https://github.com/microsoft/UFO?ref=aipure
Product Information
Updated:Jun 16, 2025
What is UFO²
UFO² (Desktop AgentOS) is an open-source project developed by Microsoft that represents a major evolution from the original UFO framework. It serves as a comprehensive multi-agent system designed to automate Windows operations through natural language commands. Released in April 2025, UFO² integrates with Windows OS (version 10 and above) and requires Python 3.10 or higher. The framework is not an official Windows feature but rather an experimental platform that demonstrates advanced automation capabilities through the combination of large language models, computer vision, and system integration.
Key Features of UFO²
UFO² (Desktop AgentOS) is an advanced UI-focused multi-agent framework for Windows OS that transforms natural language requests into automated workflows. It combines deep OS integration, hybrid GUI and API actions, and a continuous knowledge system to execute complex tasks across multiple applications. The system features picture-in-picture desktop isolation, speculative multi-action execution, and sophisticated control detection through both UI automation and visual recognition.
Deep OS Integration: Combines Windows UIA, Win32, and WinCOM for comprehensive system control, enabling both GUI automation and direct API commands
Picture-in-Picture Desktop: Runs automation tasks in an isolated virtual desktop environment, allowing users to continue working on their main screen without interference
Multi-Agent Coordination: Uses HostAgent to manage task planning and multiple AppAgents to handle application-specific operations, enabling complex cross-application workflows
Knowledge Substrate System: Integrates multiple knowledge sources including offline documentation, online search, user demonstrations, and execution traces through RAG technology
Use Cases of UFO²
Office Automation: Automates routine tasks across Microsoft Office applications, such as data entry, document formatting, and email management
System Administration: Handles complex Windows system operations and configurations through natural language commands
Cross-Application Workflows: Executes tasks that span multiple applications, such as collecting data from a webpage and creating a spreadsheet report
Custom Application Control: Can be trained to operate specialized or niche applications through demonstration learning and documentation
Pros
Higher success rate compared to traditional automation tools
Flexible combination of GUI and API-based actions
Continuous learning capability through various knowledge sources
Non-intrusive operation through virtual desktop isolation
Cons
Currently limited to Windows OS environment
Requires API keys and configuration setup
May have privacy concerns when handling sensitive information
How to Use UFO²
Install Prerequisites: Ensure you have Python >= 3.10 and Windows OS >= 10 installed on your system. Optionally create a conda environment with 'conda create -n ufo python=3.10'
Clone and Install UFO: Clone the repository with 'git clone https://github.com/microsoft/UFO.git', navigate to UFO directory, and run 'pip install -r requirements.txt'
Configure LLM Settings: Copy ufo/config/config.yaml.template to ufo/config/config.yaml and configure your LLM settings (OpenAI or Azure OpenAI) including API keys and endpoints for both HostAgent and AppAgent
Setup RAG (Optional): Configure optional Retrieval Augmented Generation (RAG) features in config.yaml - can include offline help documents, Bing search, self-experience, or user demonstrations
Launch UFO: Start UFO by running 'python -m ufo --task <your_task_name>' for interactive mode, or 'python -m ufo --task <your_task_name> -r "<your_request>"' for direct execution
Monitor Execution: Check the ./ufo/logs/<your_task_name>/ directory for execution screenshots and request/response logs to monitor or debug the agent's actions
Get Support: For help, check documentation at microsoft.github.io/UFO/, create GitHub issues, or contact [email protected] for other communications
UFO² FAQs
UFO² is a Desktop AgentOS, which is a new generation of agent framework that can run on Windows desktop OS. It's designed to turn natural-language requests into automatic, reliable, multi-application workflows on Windows, beyond UI-Focused capabilities.
UFO² Video
Popular Articles
SweetAI Chat vs HeraHaven: Find your Spicy AI Chatting App in 2025
Jul 10, 2025
SweetAI Chat vs Secret Desires: Which AI Partner Builder Is Right for You? | 2025
Jul 10, 2025
How to Create Viral AI Animal Videos in 2025: A Step-by-Step Guide
Jul 3, 2025
Top SweetAI Chat Alternatives in 2025: Best AI Girlfriend & NSFW Chat Platforms Compared
Jun 30, 2025
Latest AI Tools Similar to UFO²
Athena AI
Athena AI is a versatile AI-powered platform offering personalized study assistance, business solutions, and life coaching through features like document analysis, quiz generation, flashcards, and interactive chat capabilities.
Aguru AI
Aguru AI is an on-premises software solution that provides comprehensive monitoring, security, and optimization tools for LLM-based applications with features like behavior tracking, anomaly detection, and performance optimization.
GOAT AI
GOAT AI is an AI-powered platform that provides one-click summarization capabilities for various content types including news articles, research papers, and videos, while also offering advanced AI agent orchestration for domain-specific tasks.
GiGOS
GiGOS is an AI platform that provides access to multiple advanced language models like Gemini, GPT-4, Claude, and Grok with an intuitive interface for users to interact with and compare different AI models.
Popular AI Tools Like UFO²
ChatGPT
ChatGPT is an advanced AI-powered chatbot developed by OpenAI that uses natural language processing to engage in human-like conversations and assist with a wide range of tasks.
SearchGPT
SearchGPT is an AI-powered search prototype by OpenAI that provides fast, conversational answers with clear sources using GPT models.
OpenAI
Free TrialLarge Language Models (LLMs)
OpenAI is a leading artificial intelligence research company developing advanced AI models and technologies to benefit humanity.
Open AI o3
OpenAI o3 is a next-generation AI reasoning model with enhanced capabilities in coding, mathematics, and problem-solving that achieves breakthrough performance on advanced benchmarks while prioritizing safety through deliberative alignment.