Non, merci!

Project Overview

Non, merci! is a humorous, systemic game developed over a single weekend for the Brno Game Jam 2025. Inspired by titles like Untitled Goose Game, players take control of a mischievous dog named Merci. The core gameplay revolves around navigating a dynamic world populated by AI-controlled characters, completing relentless, multi-step quests, and causing general chaos through a rich set of canine interactions.

To accelerate our jam workflow, we utilized a pre-existing 3D dog asset with pre-made animations, as well as stylized polygon city models. However, all of the code, controllers, and game systems were built entirely from scratch by our team over the 48-hour period.

My Contributions & Technical Highlights

As a generalist on the team, I helped shape the core dog controller and interaction logic, but my primary focus was on architecting the Artificial Intelligence and integrating the Audio Pipeline.

• Modular AI System: I developed a robust, state-based AI Controller capable of handling complex NPC behaviors. The system features a Patrol mechanic with configurable waypoints, dynamic detection and chase mechanics, and a range-based combat/aggression system. By using an Action Scheduler to decouple movement from action logic, the AI smoothly transitions between states (e.g., getting annoyed, chasing the dog, cooling down, and returning to patrol).
• Professional Audio Integration (FMOD): To elevate the game’s polish beyond standard jam quality, I integrated FMOD Studio for audio management. This enabled dynamic, parameter-driven soundscapes — including varied dog vocalizations, surface-specific footsteps, and emotion-tied audio triggers for the NPCs. The audio ended up being a highly praised feature.
• Emotion & Visual Feedback: We implemented a sprite-based emotion system to give players clear, immediate feedback. NPCs display animated visual cues (driven by DOTween) that reflect their current AI state and suspicion levels.

Core Systems & Architecture

The project was built with a heavy emphasis on clean, scalable systems design, utilizing interfaces, event-driven architecture, and ScriptableObjects:

• Advanced Dog Controller: A state machine utilizing Animator parameter hashing to manage over 13 distinct dog actions. This includes combat (attack, grab-and-bite), environmental actions (dig, pee, poop), and behavioral idle animations (yawn, sleep, shake-and-dry).
• Type-Safe Interaction Framework: We built a highly extensible interaction system using abstract base classes. A strict InteractionType enum prevented invalid interactions at compile-time, allowing world objects to listen for and respond to specific canine actions.
• Event-Driven Quest Engine: Quests were structured as sequences of interactive subtasks. The event-driven design allowed dynamic quest completion and updates without relying on brittle, hardcoded logic.

Reception

The game was well-received by the Brno Game Jam community and judges. Players highly praised its charm, the hilarious interactivity of the dog’s mechanics, and how remarkably polished the interconnected systems (AI reactions, audio, and quest progression) felt for a game created in just one weekend.