MON AMI

Lore :
In my village, there’s a rumor: the old abandoned hospital is haunted. People have tried to go there, but… the few who came back looked like they’d lost their minds. One talks of monsters, another of rituals. Though their stories don’t make sense, they all agree on one thing: something was trying to keep them there. Something they call Mon Ami.

Tonight, it’s decided. My friends and I are going to explore that place! Since the entrances have been locked by the authorities, we’ll go through the underground elevator. 

This game was developed by five students during the winter semester of the “Master's in Computer Science, specializing in Video Game Development” program at the University of Sherbrooke (DDJV). 

Over the course of four months, we learned to use Unreal Engine 5.7 with the goal of creating a real-time multiplayer horror game, inspired by "Lethal Company", "R.E.P.O.", and "Escape the Backrooms". This allowed us to apply what we learned in our courses on C++, Gameplay, Back-End, Digital Media, and Artificial Intelligence.

Attached are videos showcasing the game as well as the systems we implemented throughout the project.

This project was particularly challenging, and we hope you enjoy playing it as much as we enjoyed creating it.

Inputs : 

• WASD : Movement
• Run : Shift
• Crouch : CTRL
• Jump : Space bar
• Interact : E
• Use item : left click
• Throw item : Q
• Change item in inventory : Scroll
• Pause : Escape 

Systems :

Procedural Dungeon

• Rooms are generated from data-driven sets with weighted probabilities that evolve over the run
• Ensures varied layouts while keeping pacing and structure under control

Spawning System

• Actors (items, enemies, props) are spawned using configurable probabilities and weighted selection
• Supports custom spawn strategies for different gameplay needs

Dynamic Quest System

• Objectives are generated based on the current dungeon content
• Automatically selects required and optional goals, including collectible quotas

Adaptive Difficulty

• Difficulty scales based on player performance between runs
• Uses curve-based evaluation to adjust dungeon size, objectives, and intensity

Asymmetric Perception

• Players experience different versions of the world (visuals, sounds, interactions, and monsters)
• Implemented via network-level filtering so each client only receives relevant data, with a special attention to Navmesh

AI Director

• Enemy spawning is driven by player “stress” levels
• Dynamically shifts between calm, buildup, and high-intensity phases

Dynamic Music System

• Music adapts in real time to player state (stress, movement)
• Layered audio system that activates/deactivates tracks based on gameplay

Interaction System

• Unified interaction model (press / hold) based on player focus and proximity
• Server-authoritative logic ensures consistent multiplayer behavior

Inventory & Abilities

• Items grant abilities using a modular ability system
• Supports consumables, equipment, and toggleable effects

Quest System

• Players must complete objectives and regroup to progress
• Special events can temporarily gate progression and create tension
  

Credits : 

• Elias "Xyro" Del Pozo : Monsters, AI Director, Adaptative Difficulty, and rendering
• Thomas Vanwalleghem : Monsters, Dynamic Music System, Level Design and Gameplay
• Florent Gilloteaux : UI / UX, Inventory, Items, and GAS
• Sylvain Brunet : Back-End, Asymmetric perception, Spawning system, and Level Design
• Nell Truong : Back-End, Interactions, Quests, Dungeon generation, and GAS  
• Alexandre Roche : Composer, Sound

Special thanks to Claire Flambeau for the Scooby Doo inspired plush toy asset and the help in UX design.

The rest of the graphic assets can be found on FAB.