Design
DESIGN
THE
SYSTEM.
An open design framework for AI-native products. Built around human intent, not just interfaces.
Explore the frameworkTHE INTERFACE
IS NOT THE
AI products do more than display information. They interpret, reason, decide, act, and learn.
TUN defines the relationship between humans, AI, agents, data, tools, decisions, actions, and learning. A shared language for designing intelligent systems people can understand and control.
Read the concept noteFROM INTENT
TO
A simple structure for complex intelligence. Each stage makes the system more legible and accountable.
Ask
Understand intent.
Think
Interpret context.
Propose
Make choices clear.
Act
Execute with authority.
Learn
Improve from outcomes.
BUILT ON
A system should remain understandable and under human direction, even as its intelligence grows.
Transparent
Make actions, information, and intent understandable.
User sovereign
Keep meaningful human control in the loop.
Natural
Let people express intent in their own language.
Systemic
Design relationships, not isolated screens.
Adaptive
Respond to context without becoming unpredictable.
Reversible
Make important actions inspectable and correctable.
Composable
Build with interoperable parts and clear boundaries.
Calm
Surface what matters. Keep complexity underneath.
INTENT,
MADE
A shared visual language for intelligent products.
Light and dark themes, semantic AI states, and explicit controls translate the specification into a working token package. Color describes a state; it never grants authority or proves an outcome.
142 declared contrast pairs checked. Token-level validation only — not an accessibility certification. The browser specimen simulates its approval flow; it does not publish or save data.
EXPLORE
THE
The thinking, rules, and building blocks behind TUN. Open to inspect, use, and help shape.