UI-TARS-desktop

The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra

23 articles 31.6k View on GitHub ↗
23 articles
StatusEnum States in the UI-TARS Agent Lifecycle: Complete Reference

Explore the eight StatusEnum states like INIT, RUNNING, and ERROR that define the UI-TARS Agent lifecycle from start to finish. Understand agent behavior and transitions.

api-reference
May 10, 2026
How UI-TARS Handles User‑Initiated Stops During Agent Execution

Learn how UI-TARS desktop handles user initiated stops. The agent sets a stopped flag in AgentContext to break the execution loop and emit a TASK_CANCEL event.

internals
May 10, 2026
UITarsModelVersion and Provider Selection Relationship in UI-TARS Desktop

Understand the UITarsModelVersion and provider selection relationship in UI-TARS Desktop. Learn how VLMProviderV2 maps to model versions with fallback options.

deep-dive
May 10, 2026
How to Set Up a Custom Report Storage Server for HTML Exports in UI-TARS

Learn how to set up a custom Report Storage Server for UI-TARS HTML exports. Implement a POST endpoint to store reports and activate automated uploads via Report Settings.

how-to-guide
May 10, 2026
UI-TARS Coordinate Transformation: 3 Factors Driving Normalization Calculations

Discover the 3 key factors driving UI-TARS coordinate transformation calculations: raw input, NormalizeCoordinates function, and screenshot dimensions for accurate scaling.

deep-dive
May 10, 2026
How the Model Invoke Retry Mechanism Handles API Errors in UI-TARS Desktop

Learn how the UI-TARS desktop SDK uses async-retry for its model invoke retry mechanism to handle API errors, automatically retrying failed requests and distinguishing transient errors from user aborts.

how-to-guide
May 10, 2026
How to Implement Custom System Prompts for Specific UI-TARS Use Cases

Learn to implement custom system prompts for UI-TARS use cases. Customize GUIAgent, Operators, or SYSTEM_PROMPT_TEMPLATE for tailored interactions.

how-to-guide
May 10, 2026
How Screenshot Retry Logic Works in UI-TARS Desktop: Configuration and Customization Guide

Learn how UI-TARS Desktop's screenshot retry logic handles failures with async-retry. Customize retry attempts and callbacks using the RetryConfig interface for robust UI automation.

how-to-guide
May 10, 2026
How the Browser Operator Handles Search Engine Configurations in UI-TARS

Discover how UI-TARS desktop handles search engine configurations. Learn how DefaultBrowserOperator and RemoteBrowserOperator manage different search engines like Google Bing and Baidu.

how-to-guide
May 10, 2026
UI-TARS Loop Interval Parameter: Configuration and Optimization Guide

Optimize your UI-TARS agent with the loopIntervalInMs parameter. Learn how to configure this crucial setting to balance automation speed and system stability for peak performance.

how-to-guide
May 10, 2026
How to Configure the Responses API for Reduced Token Consumption in UI-TARS

Reduce token consumption with UI-TARS Responses API. Enable useResponsesApi: true and set max_tokens for efficient message handling and automatic pruning.

performance
May 10, 2026
Error Status Types in UI-TARS: Complete Handling Guide for GUI Agent Failures

Master UI-TARS error status types with our complete handling guide. Learn to manage GUI agent failures using ErrorStatusEnum and GUIAgentError for robust recovery.

how-to-guide
May 10, 2026

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"
Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →