CaseZero

A complete visual concept for The Missing Prototype, showcasing the player journey from starting an investigation to examining clues etc.
A cinematic, dark-themed detective mystery game UI featuring the case dashboard, suspect profiles etc.

Inspiration

I wanted to explore how AI could make mystery games more interactive and engaging. Solving a detective case involves more than finding clues: it requires questioning suspects, connecting evidence, spotting contradictions, and building a convincing theory. This inspired me to create CaseZero, a detective mystery game that combines investigative gameplay with agentic AI.

What it does

CaseZero is a detective mystery game built around The Missing Prototype, a fictional case involving a research prototype that disappears from a secured facility.

Players investigate the incident, interview three suspects, collect evidence, reconstruct the timeline, develop theories, and make a final accusation. Four specialized AI agents are planned to assist with investigative decisions, suspect interviews, evidence analysis, and challenging the player's theories.

The goal is to help players think like detectives instead of simply asking a chatbot for the answer.

How we built it

CaseZero is being developed using Jac, with a focus on separating deterministic game logic from AI reasoning. The backend is designed to manage case facts, investigation sessions, evidence, and verdicts, while specialized AI agents assist with different parts of the investigation.

The interface uses a simple detective-inspired design with dark backgrounds, readable text, and subtle gold accents. We are also focusing on session management, secure API credentials, input validation, and reliable fallback behavior when AI is unavailable.

Challenges we ran into

One major challenge is keeping investigation sessions consistent across different screens and backend requests. Suspect profiles, evidence, and timeline events must all belong to the correct active investigation.

Another challenge is making AI useful without allowing it to invent clues or change the mystery's solution. The game needs a reliable source of truth so that AI-generated dialogue and analysis remain consistent with the case.

Accomplishments that we're proud of

We are proud of the concept of combining detective gameplay with specialized AI agents that have distinct responsibilities. Rather than relying on one chatbot, CaseZero is designed to support investigation, questioning, evidence analysis, and critical thinking through separate agent roles.

We are also working toward a maintainable architecture where the game remains playable even when an AI service is unavailable.

What we learned

This project has helped us understand the importance of backend reliability, session persistence, secure API integration, and clear boundaries between AI-generated reasoning and deterministic application logic. We have also learned that an engaging AI application requires more than a good interface: the underlying game mechanics must work consistently.

What's next for CaseZero

The next steps are to finish and test the investigation flow, connect the suspect, evidence, and timeline systems, integrate the AI agents, and verify the complete mystery from start to finish. Future improvements could include additional cases, more complex evidence chains, richer suspect conversations, and new investigative challenges.