Full-stack AI game project note
Artemis Lost
Full-stack lunar incident simulator with an OpenAI-powered mission director, deterministic local game systems, role-based crew play, and durable saves.
Project Note
Artemis Lost is a full-stack sci-fi command game where an OpenAI-powered mission director reacts to a stranded lunar crew in crisis. It turns structured AI interaction into a playable, stateful, role-based mission experience instead of a generic chat interface.
The project began under hackathon conditions and became a solo build almost immediately. That context shaped the work: the system had to feel memorable and alive, but disciplined enough not to collapse into loosely connected AI features.
What I Built
- React and Vite frontend for crew creation, launch flow, live mission UI, and win/loss resolution screens.
- Node and Express backend for mission-director turns, prompt assembly, and session handling.
- OpenAI Responses API integration for narration and structured mission consequences.
- Deterministic local gameplay systems for crew roles, mission mechanics, handoffs, and outcome resolution.
- Durable save support through browser-scoped slots and Postgres-backed persistence.
- Testing coverage across frontend behavior and backend routes.
Public Links
Interface Evidence
The Devpost screenshots are useful because they show the project as an actual playable interface rather than only a codebase. They make the core loop visible: save slots, crew creation, launch transition, live mission state, and mission resolution.
Why It Matters
The central design problem was keeping AI narration useful without giving it total authority over game state. The mission director returns narration and structured partial updates, while local systems preserve role logic, turn flow, crew coordination, mission pressure, and outcome resolution.
That hybrid architecture makes the game stronger evidence than a simple AI wrapper. It shows full-stack implementation, game design judgment, prompt integration, state management, and interface work under time pressure.