AgentIndex icon
AgentIndex
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Submit Tool
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Home/
Vision / Multimodal/
UI-TARS-desktop
UI-TARS-desktop logo

UI-TARS-desktop

Active·★ 39.6k·Apache-2.0·Updated 2026-10-09
★ Most Popular★ Essential

This ByteDance repository hosts TARS, an open-source multimodal AI agent stack with two projects. Agent TARS is a general multimodal agent with a CLI and web UI that combines GUI-agent vision with DOM-based browser control and MCP tools, aiming to complete tasks the way a person would. UI-TARS Desktop is a native desktop application that lets a vision-language model, driven by the UI-TARS and Seed-VL series models, operate your local computer, a remote computer or a browser with mouse and keyboard.

It suits developers and researchers exploring computer-use agents: automating GUI tasks in desktop apps that have no API, multi-step browser tasks such as comparing and booking travel, or prototyping agents that see the screen. UI-TARS Desktop works on Windows and macOS and in the browser, with screenshot-based recognition and real-time status feedback; Agent TARS runs headful in a web UI or headless as a server and can mount additional MCP servers.

Getting started: run Agent TARS with `npx @agent-tars/cli@latest` (Node.js 22+ for a global install) and pass your model provider and API key, for example Volcengine or Anthropic. UI-TARS Desktop is a separate desktop app; follow its quick start to download it and configure a vision-language model endpoint.

This write-up is based on the project's README and public repository data. We have not hand-tested the tool. Data as of 2026-11-07.

License
Apache-2.0
Latest release
v0.3.1 · 2026-10-08
Last commit
2026-10-09
Contributors
50
Forks
4,027
Open issues and PRs
511
First published
2025-01-19

Source: GitHub, npm and PyPI public APIs.

#GUI Agent#Desktop App#Multimodal AI#Automation#Cross-Platform
$ Install
$ npx @agent-tars/cli@latest
↗ Visit site★ GitHub
01

Features

01Native GUI agent that sees the screen and interacts with desktop applications
02Multimodal LLM-powered visual understanding of any UI
03Browser automation and shell command execution built in
04MCP integration for extending agent capabilities with custom tools
05Cross-platform desktop app with web UI option
02

Why choose it

+Operates any GUI through vision, including desktop apps without APIs.
+Two complementary tools: a desktop GUI agent and a general multimodal agent with CLI, web UI and MCP support.
+Supports local, remote computer and browser operators.
+Backed by ByteDance's UI-TARS research models; Apache-2.0 licensed.
03

Trade-offs

!GUI agents act directly on your screen with your permissions; supervise them and avoid running them on machines with sensitive sessions.
!Vision-based control is slower and less deterministic than API or DOM automation, and results depend on the model you use.
!You need access to a capable vision-language model, either hosted or self-deployed, which can be costly to run.
05

Maintenance check

ActiveChecked: 2026-11-06
Last commit
2026-10-09
Commits, last 90 days
8
Releases, last 90 days
1
Latest release
v0.3.1 (2026-10-08)
Open issues
511
Recent issues with at least one comment
13/16
Commits per week, last 12 weeks
Commits per week, last 12 weeks: 0, 0, 0, 0, 0, 5, 0, 3, 0, 0, 0, 0

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

  • #268 [Feature] Windows platform support for Agent TARS 23 comments · 2025-03-22
  • #591 [Bug]: Missing clicks20 comments · 2025-05-13
  • #587 [Bug]: UI-TARS-2B-SFT model maximum context length Error16 comments · 2025-05-13
  • #11 no action after thinking12 comments · 2025-01-22
  • #311 Mac已经按要求授权应用访问设备的权限,依然无法执行浏览器访问11 comments · 2025-03-24

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

04

Who it fits

Pick it if
  • You want a visual GUI agent that understands native desktop screens and operates mouse/keyboard actions end-to-end via vision models.
  • You want an open-source multimodal desktop application for cross-application task automation on macOS, Windows, and Linux.
Look elsewhere if
  • You only need web-browser-based automation; Playwright or Scrapling are much faster and more reliable via direct DOM.
  • Your environment lacks a capable GPU or reliable multimodal LLM API with high vision resolution support.
06

How it differs

Open Interpreter logovs Open Interpreter★ 68.9k

UI-TARS Desktop operates through visual GUI perception, mouse movements, and screen recognition. Open Interpreter operates primarily through the host CLI and executing generated Python, Bash, or AppleScript commands.

Full comparison →
07

Compatibility

macOS
Supported
Verified via docs
Windows
Supported
Verified via docs
Linux
Supported
Verified via docs
08

Quick start

1
$ npx @agent-tars/cli@latest
09

Use cases

↳Automating desktop GUI workflows that don't have APIs by seeing and clicking the UI
↳Building multimodal agents that combine screen understanding with web and file operations
↳Running AI-assisted computer tasks through natural language instructions on desktop
10

Alternatives

⚡Top VS:
VSOpen Interpreter★68.9k→VSvideo-search-and-summarization★2.0k→VSMeshroom★13.1k→VSGemini CLI★108.9k→
Open Interpreter logo
Open Interpreter★ 68.9k
A coding agent for the terminal, optimized for low-cost and open models. Forked from OpenAI's Codex, it can emulate other agent harnesses and runs commands in a native sandbox.
VSCompare→
video-search-and-summarization logo
video-search-and-summarization★ 2.0k
NVIDIA AI Blueprint for video search and summarization (VSS) is a GPU-accelerated reference architecture for building video analytics agents with real-time verified alerts, visual Q&A, and automated reporting. The VSS Blueprint uses vision language models (VLMs) such as NVIDIA Cosmos, LLMs such as NVIDIA Nemotron, RAG, and NVIDIA NIMs.
VSCompare→
Meshroom logo
Meshroom★ 13.1k
Node-based Visual Programming Toolbox
VSCompare→
Gemini CLI logo
Gemini CLI★ 108.9k
An open-source AI agent that brings the power of Gemini directly into your terminal. Supports native MCP.
VSCompare→
Cherry Studio logo
Cherry Studio★ 53.6k
A powerful desktop client for multiple LLMs. Supports local and cloud models.
VSCompare→
GitHub MCP Server logo
GitHub MCP Server★ 34.5k
GitHub's official MCP Server. Allows AI agents to interact directly with your GitHub repositories (read files, search code, issues).
VSCompare→
budibase logo
budibase★ 28.4k
Create business apps and automate workflows in minutes. Supports PostgreSQL, MySQL, MariaDB, MSSQL, MongoDB, Rest API, Docker, K8s, and more 🚀 No code / Low code platform..
VSCompare→
keep logo
keep★ 12.5k
The open-source AIOps and alert management platform
VSCompare→
See all alternatives →

Comments

Log in to leave a comment

No comments yet. Be the first!

On this page
01Features02Why choose it03Trade-offs05Maintenance check04Who it fits06How it differs07Compatibility08Quick start09Use cases10Alternatives
bytedance /
UI-TARS-desktop
↗

ByteDance's multimodal agent stack and desktop GUI agent

39.6k4.0kApache-2.0
Last commit1mo ago
StatusActive
CategoryVision / Multimodal
Contributors50
Latest releasev0.3.1
Links
Documentation↗Discussion↗Issues↗Releases↗
Cloud Infrastructure
Ad

Deploy on DigitalOcean

Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.

Claim $200 Free Credit
© 2026 AgentIndex.app|Built by a 10-year iOS Developer.
Best by use caseAboutContactAdvertisePrivacy PolicyTerms of UseQYSGitHubBuy me a coffee ☕

Browse by Category

MCP ServersCode AssistantWorkflow AutomationRAG / Knowledge BaseMulti-AgentBrowser AutomationLLM InfraDev ToolingObservability

Not affiliated with Anthropic, OpenAI or Microsoft.