UI-TARS Desktop is a GUI Agent application based on UI-TARS (Vision-Language Model) that allows you to control your computer using natural language. It integrates key components like perception, reasoning, grounding, and memory into a single vision-language model, enabling end-to-end task automation without predefined workflows or manual rules.
Product Screenshot
Features
Comprehensive GUI Understanding
Processes multimodal inputs (text, images, interactions) to build a coherent understanding of interfaces, with real-time monitoring and accurate response to dynamic GUI changes
Unified Action Space
Standardized action definitions across platforms (desktop, mobile, and web), supporting additional operations like hotkeys, long press, and platform-specific gestures
Dual System Reasoning
Combines fast intuitive responses with deliberate high-level planning, supporting multi-step planning, reflection, and error correction for robust task execution
Memory System
Includes short-term memory for capturing task-specific context and long-term memory for retaining historical interactions and knowledge to improve decision-making
Pricing
Open Source Version
Free
- Apache 2.0 license
- Local deployment support
- Cloud deployment support
- Multiple model sizes available (2B, 7B, 72B)
- Requires self-configuration of API
Use Cases
Automated Task Execution
Uses natural language instructions to automatically execute various computer tasks, such as browsing websites and sending tweets
Cross-Platform Operation
Supports automated operations across different platforms (Windows, MacOS), providing a unified user experience