// Drive browsers & computers
UI-TARS-desktopv0.3.1
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
Open source Maintained For agents
IN ONE PARAGRAPH
UI-TARS-desktop is the Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra.
Good for: Developers who want their agent (Claude Code, Codex, Cursor…) to drive browsers & computers.
01How to start
- 1Read the official README for requirements and permissions.
- 2Install it as the official guide describes.
- 3Tell your agent the command exists (e.g. in AGENTS.md / CLAUDE.md) so it knows when to call it.
02Alternatives & comparison
Compare side by side → mobile-mcp Model Context Protocol Server for Mobile Automation and Scraping (iOS, Android, Emulators, Simulators and Real Devices) ★ 9.9k nanobrowser Open-Source Chrome extension for AI-powered web automation. Run multi-agent workflows using your own LLM API key. Alternative to OpenAI Operator. ★ 14k browser-use Agents that use the browser. ★ 119k chrome-devtools-mcp Chrome DevTools for coding agents ★ 54k playwright-mcp Playwright MCP server ★ 38k
03Related
- ›Hugging Face AI Agents Course · Hugging FaceMultilingual (incl. EN/ZH)
- ›huggingface/agents-course · Hugging FaceEnglish
- ›AI Agents for Beginners · MicrosoftMultilingual (incl. EN/ZH)
- ›Building effective agents · AnthropicEnglish
- ›OpenAI: A practical guide to building agents · OpenAIEnglish
- All tutorials →
Versions, why it is listed, for AI to read
Recent releases
- v0.3.1 2026-10-08
- v0.3.0-canary-263cff2e9-20251120130823 2025-11-20
- v0.3.0-canary-3262c6416-20251106135418 2025-11-06
- v0.3.0 2025-11-04
- v0.3.0-beta.11 2025-09-09
For AI agents
- Interfaces：App · CLI
- Machine-readable record：/api/v1/software/ui-tars-desktop.json · llms.txt
- Added 2026-10-10 · updated 2026-10-22