Orient UI is a design system for AI conversation. It gives Claude (and other AI tools) persistent orientation, adaptive visual rendering, human-in-the-loop checkpoints, and structured surfaces.

It turns unstructured AI conversations into guided interactions — interviews, diagnostics, decision support, design sessions, and elicitation flows that keep you and the AI oriented throughout.

Orient UI.md — full canonical protocol with all surfaces, governance patterns, and visual grammar

Orient UI Skill.md (Claude format) — drop it into Claude and Orient UI activates immediately

Orient UI YAML.md — structured YAML representation for agent configuration, routing, and implementation specs

Orient UI TOON.md — token-optimized object notation for compact runtime context

README.md — navigation and usage guide for deciding which format to use and how to apply the system

Designers, consultants, operators, researchers, and anyone running structured conversations with AI who wants better pacing, clearer state tracking, and human-in-the-loop control over meaning and decisions.

Persistent orientation, progressive expansion

One-question rule (one active question per turn, always)

Human-in-the-loop protocol (AI never finalizes meaning without confirmation)

Smallest useful rendering (escalate structure only when it helps cognition)

Named surfaces for interviews, scorecards, decision reports, wireframes, diagrams, charts, personas, empathy maps, and hybrid views

Orient UI is the same system I use to shape context before every design engagement.