Revna AI
Revna builds action intelligence for game worlds, enabling characters to understand intent and take physical, strategic, and social action within a game's physics and rules. The company develops physics-based agents that translate language into coordinated full-body movement and strategic agents for multiplayer board games that plan across turns. Revna operates as both an AI research lab and a game studio, grounding its technology in playable experiences.
- Artificial Intelligence
- Gaming & eSports
- AI Agents
- Software Only
Funding
Founders
Product
Problem
Most AI can generate language or content, but game characters still have narrow, scripted ways to act, limited to a small set of predefined responses. This missing layer of action intelligence prevents characters from understanding intent and adapting their behavior to the changing state of the world, other agents, and the constraints of their own bodies.
Solution
Revna builds action intelligence that connects intent to behavior, enabling agents to understand what a person wants, perceive their situation, choose what to do, and carry it out within the constraints of a body, a game, and other agents. The company is developing two complementary forms of action intelligence: physics-based agents that translate language into coordinated full-body movement accounting for balance, contact, and momentum, and strategic agents for multiplayer board games that understand rules, plan across turns, and adapt to other players' decisions. Revna operates as both an AI research lab and a game studio, with researchers, engineers, designers, and producers working together from the beginning to identify experiences that require new forms of intelligence and build playable games around them. The longer-term vision is to create agents that pursue goals, respond to other characters, make decisions, and produce lasting effects within increasingly complex game worlds.
Target Audience
Primary customers are game developers and studios seeking AI-native game mechanics that enable expressive character movement, adaptive decision-making, and coordination that cannot be achieved with traditional scripted behavior systems.
Features
- Physics-based agents that translate language into coordinated full-body movement, accounting for balance, contact, momentum, and changing environmental state rather than selecting from a fixed animation library
- Strategic agents for multiplayer board and turn-based games that understand rules, evaluate possible actions, plan across turns, and adapt to other players' decisions
- Goal of strategic intelligence that transfers across a family of games rather than requiring a separate hard-coded bot for every title
- Integrated research-and-studio workflow where real games expose agents to varied interactions and unexpected situations, generating new research questions
- Development pipeline spanning four stages: bodies (physical control), strategies (rule-based decision-making), teams, and worlds (persistent, social, adaptive NPCs)