Shadow Verdict: A Detective Game

The Loading Screen
AI Analysis - to find relationships between know evidence
Evidence from the crime scene
Evidence Board
Autopsy Report
Crime Scene
Interrogation Chat
Final Accusaation
This is an example output

Inspiration

I’ve always been fascinated by psychological thrillers and detective stories where the truth feels uncertain and every character hides something beneath the surface. I wanted to create an experience where players don’t just solve a mystery but truly feel like investigators: questioning suspects, analyzing contradictions, connecting clues, and dealing with manipulation, fear, and deception.

Most detective games follow fixed dialogue trees and predictable investigation systems. I wanted Shadow Verdict: A Detective Game to feel more alive and psychologically immersive, where each suspect behaves differently, evidence evolves dynamically, and players must think critically instead of simply clicking through dialogue.

That idea became the foundation of Shadow Verdict: an AI-powered noir detective experience focused on deduction, atmosphere, and psychological tension.

How I Built It

I built the entire project using MeDo, leveraging AI-assisted full-stack development through natural language prompting and iterative refinement.

The project began as a simple interrogation-based detective concept but gradually evolved into a much more immersive investigative experience featuring:

AI-driven suspect interrogation system
Dynamic suspect personalities and emotional behaviors
Evidence discovery and investigation mechanics
Crime scene exploration
Autopsy report system
AI Analysis Board for clue correlation
Cinematic noir-inspired UI/UX
Interactive tutorial system with a detective mascot
Layered ambient sound design and cinematic transitions

The investigation system was designed to feel interactive and psychologically engaging rather than linear. Players can:

Explore crime scenes
Question suspects
Unlock forensic findings
Analyze correlations between clues
Build their own deductions before making a final accusation

I continuously refined the application through multi-turn development iterations, focusing heavily on:

immersion
UI polish
realism
emotional behavior
performance optimization
investigation flow
readability and user experience

The app was deployed directly using MeDo’s built-in deployment tools.

Challenges I Faced

Building Realistic Interrogations

One of the biggest challenges was making suspects feel believable. I wanted different personalities such as manipulative, anxious, calm, or aggressive suspects to react differently during conversations without becoming repetitive or predictable.

Dynamic Evidence System

Creating a clean evidence system was surprisingly difficult. Evidence needed to:

unlock naturally
avoid duplication
stay readable
connect logically with the investigation

I also redesigned the evidence experience multiple times to reduce information overload and improve immersion.

Performance & Stability

As more cinematic systems were added (including audio, particles, overlays, transitions, portraits, and tutorials), deployment stability became a major challenge. Several systems worked in preview mode but caused issues in production builds, requiring extensive debugging and optimization.

Immersion & UX Design

A major realization during development was that immersion can break very easily. Small issues, like:

loading delays
overwhelming text
inconsistent portraits
broken transitions
repetitive UI

could completely affect the detective atmosphere. A large part of development became focused on polishing the experience rather than simply adding features.

What I Learned

This project taught me that building an engaging product is about much more than functionality.

Some of the biggest lessons I learned were the following:

Strong UX design is just as important as core features
Small polish improvements can dramatically improve immersion
AI systems need fallback logic and error handling
Psychological engagement matters more than visual complexity
Iterative refinement is essential for building polished experiences

I also learned how powerful AI-assisted development tools like MeDo can be when combined with creative iteration and strong product thinking.

Most importantly, I learned how to design around player experience: not just what the app does, but how the player feels while using it.

What’s Next

I plan to continue expanding Shadow Verdict by adding:

more advanced psychological suspect behavior
branching case outcomes
richer cinematic visuals
deeper forensic systems
additional crime scenarios
improved AI deduction systems
enhanced animations and environmental storytelling

The long-term goal is to make the experience feel closer to a real interactive psychological detective thriller rather than a traditional chat-based application.