Clone Labs
Category: AI Agents
AI startup developing a 'User Model' that learns computer and agent usage patterns to predict and execute next actions on behalf of users, eliminating the human-in-the-loop bottleneck in AI agent workflows. Clone Labs was founded in 2026. The company is led by Minchan Kim (김민찬). Based in Seongnam, South Korea. Team size: 1-10. Latest round: Seed. Key investors include NAVER D2SF.
- Founded
- 2026
- Headquarters
- Seongnam, South Korea
- Team size
- 1-10
Value proposition
Eliminates the human-in-the-loop bottleneck in AI agent usage by predicting user intent and automating decision-making without requiring constant user intervention.
Products and solutions
Clone Desktop (AI builder desktop app); Clone Plugin (browser/IDE plugin for AI builders); User Model (3-layer AI: Recording → Memory → Prediction)
Unique value
User Model that learns computer and agent usage patterns to predict and autonomously execute next actions, with a confidence-based escalation system — high-confidence tasks auto-executed, low-confidence tasks sent for user confirmation.
Target customer
AI builders and developers who heavily use AI agents and face productivity bottlenecks from constant human-in-the-loop feedback requirements.
Industries served
AI Agent Infrastructure; Enterprise Productivity; Developer Tools
Technology advantage
Three-layer User Model architecture (Recording, Memory, Prediction); Computer-Use Agent (CUA) capability; Long-term memory and privacy-preserving memory architecture; Confidence-based auto-execution vs. user confirmation system; 7 published AI agent research papers including collaborations with Stanford and Carnegie Mellon University
How they differentiate
Focuses specifically on the user-model layer that predicts human intent rather than building general-purpose AI agents. The three-stage architecture (Recording → Memory → Prediction) with confidence-based auto-execution is unique. Strong academic research foundation with 7 papers from top universities.
Main competitors
Adept AI (AI agents for enterprise workflows); Cognition AI (Devin, AI coding agent); Anthropic (Computer Use feature in Claude)
Key partnerships
NAVER D2SF (investor & incubator); Stanford University (research collaboration); Carnegie Mellon University (research collaboration)
Major milestones
Discovered via NAVER D2SF Campus Tech Startup Competition (H2 2025); Incubation began January 2026; Seed investment from NAVER D2SF (May 2026); Launched Clone Desktop and Clone Plugin products; Published 7 AI agent research papers with Stanford and CMU collaborations
Market positioning
Early-stage Korean AI startup focused on the emerging 'human-agent bottleneck' problem, differentiated by its academic research pedigree (SNU, Stanford, CMU) and NAVER D2SF backing. Competes in the AI agent infrastructure layer rather than building general-purpose agents.
Geographic focus
South Korea (primary); global AI builder market (aspirational)
About Minchan Kim (김민찬)
Undergraduate student at Seoul National University (Computer Science); Visiting researcher at SNU AI Lab; Published AI agent research papers including collaborations with Stanford University and Carnegie Mellon University
Official website: https://clonelabs.org