UI-TARS logo

UI-TARS

AI AgentVerified90% conf

Open-source multimodal AI agent that performs diverse tasks in virtual environments using vision-language models

seed-tars.com

📍 Beijing, China

Verified Data

💰
Est. Revenue<$100K ARR

Open source research project from ByteDance with no monetization model - generates no direct revenue

🚀
FundingCorporate research project (ByteDance: $9.4B+ total funding, $225B valuation)

Part of ByteDance (Total funding: $9.4B+, Valuation: $225B+)

📊
Monthly Traffic500K-1M visits/month (seed.bytedance.com)

seed.bytedance.com: ~500k-1M (estimated domain traffic), github.com/bytedance/UI-TARS: High engagement (~32k stars)

👥
Users32,000+ GitHub stars
🔗github.com
🧑‍💻
Team Size20-50 employees (ByteDance Seed team)

Estimated 20-50 (ByteDance Seed team)

📈
Growth31,000 stars in first year, 500+ stars daily

Gained ~31,000 stars in its first year; currently gaining ~500+ stars daily (May 2026)

🏷️
StageCorporate Research Project
📅
Founded2024

Company Profile

ModelOpen Source Research
VerticalAI/ML, Developer tools, Automation
ClientsOpen Source Community, AI Researchers
BuyersGlobal developers, AI researchers, and engineers interested in GUI automation and AI agents
PricingFree (Open Source, Apache 2.0 License)

Contact

Strategic Analysis

Strategy

ByteDance is positioning UI-TARS as a strategic open-source research project to advance AI agent capabilities and establish thought leadership in GUI automation. The project serves as both a research vehicle and a way to attract top AI talent to ByteDance's ecosystem.

Tactics

Released as open-source under Apache 2.0 license to maximize adoption and community contributions. Leveraging GitHub for distribution and community building, with academic paper publication to establish credibility. Using ByteDance's Seed platform for official documentation and releases.

Competitive Positioning

Competes with other GUI automation tools and AI agents by focusing on native multimodal understanding and human-like interaction patterns. Differentiates through ByteDance's AI research capabilities and integration of vision-language models with reinforcement learning.

Marketing Approach

Academic-first approach with research paper publication and open-source community engagement. Leveraging ByteDance's brand recognition and developer relations through GitHub and technical conferences. Building credibility through transparent research and reproducible results.

Notable

Part of ByteDance's AI research division, gained 32K+ GitHub stars rapidly

🔗 Source ↗

Tech Stack

PythonPythonPyTorchPyTorchTransformersVision-Language ModelsReinforcement Learning
🔗 Source ↗

Recent News

Related AI Agent Companies

Discovery Sources

bytedance/UI-TARS
May 12, 2026

Signals

web traffic500K-1M visits/month (seed.bytedance.com)

seed.bytedance.com: ~500k-1M (estimated domain traffic), github.com/bytedance/UI-TARS: High engagement (~32k stars)

growth rate31,000 stars in first year, 500+ stars daily

Gained ~31,000 stars in its first year; currently gaining ~500+ stars daily (May 2026)

team size20-50 employees (ByteDance Seed team)

Estimated 20-50 (ByteDance Seed team)

user count32,000+ GitHub stars🔗 source ↗
funding raisedCorporate research project (ByteDance: $9.4B+ total funding, $225B valuation)

Part of ByteDance (Total funding: $9.4B+, Valuation: $225B+)

trend indicatorAgent🔗 source ↗
trend indicatorOpen Source🔗 source ↗
trend indicatorMultimodal🔗 source ↗
trend indicatorAI🔗 source ↗

Evidence

github.com

We’re excited to announce the release the UI-TARS-2, a major upgrade.

github.com

32,000+ GitHub stars