OUR MISSION

AI for Socially Intelligent Robots

We build the social intelligence that lets people instruct, guide, and collaborate with robots naturally — through speech, gesture, and shared understanding.

We want robots that don’t just perform useful actions, but understand intent, ask clarifying questions, and adapt as situations change.

WHAT WE ARE BUILDING

ludi

The social robot foundation model, built on real data.

Ludi is our first-generation foundation model for social interaction — trained on proprietary human and robot interaction data.

The missing layer

Today's robots can act, and conversational AI can interact — but rarely as one coherent intelligence. We believe the next leap is to integrate social understanding, communication, and physical action.

HUMAN

Speech
Gesture

ludi

Action
Speech

ROBOT

One interactive loop —
listening, clarifying, speaking, and acting in the same exchange.

RESEARCH & NEWS

Selected research & news from our labs.

Our progress — model releases, benchmark results, and technical notes from our researchers in Palo Alto and Seoul. For collaboration opportunities and inquiries about our research, please contact us at research@ludorobotics.ai

Ludi₀.₁: An Agentic System for Socially Intelligent Robots

Coming soon
Agentic AI Summit 2026, UC Berkeley, Startup Spotlight Presentation
August 2026 | Video
Prompt-to-Policy: Agentic Engineering for Reinforcement Learning
April 2026 | Github | Summary
contact@ludorobotics.ai copied to clipboard