• 

  The Logo for the Game

• 

  Godot 4 editor

• GIF

  combining ingredients to create medicine

• 

  NPC Talks about something

• 

  Npc speaks about a certain topic

• 

  NPC speaks about his disease

• 

  provide medicine to the NPC

• 

  our main base

• 

  checking the inventory and picking fruits

• 

  creating the medicine, inside the hospital

• 

  collecting fruits and vegetables grown at a farm

• 

  town 1 ( silverspine) view from godot editor

• 

  (Forest) view from godot editor

• 

  town 2 (Ironfall) view from godot editor

Inspiration

We set out to create a game that integrates Generative AI at its core, blending entertainment with real-world applications. Drawing inspiration from RPGs and shooting games, we envisioned a project that combines engaging gameplay with meaningful interactions. We aim to create a game that entertains, educates, and immerses players in a world driven by Generative AI.

What It Does

Our game employs a farming-based approach where players:

• Collect plants, battle monsters, and craft unique medicines from their harvest.
• Begin their journey with a brief, immersive introduction to the world's story and context.
• Experience dynamic NPC conversations powered by Generative AI, which discuss game lore, environmental benefits, and health-related topics. This makes our NPCs stand out from traditional games, offering more engaging and informative interactions.

How We Built It

We utilized a range of powerful tools and technologies:

• Amazon Q for analytics and backend support.
• Godot Engine for game development.
• AWS Lambda and AWS Bedrock (Mistral model) to manage NPC behavior.
• SUNO AI to generate immersive background music.
• Eleven Labs for high-quality voiceovers in the game intro.
• FLUX-dev model to produce stunning visuals for the game's introduction.

Challenges We Faced

The development process had its hurdles:

• Steep Learning Curve: As first-time users of Godot and AWS services, we needed to learn and adapt quickly.
• AWS Lambda Configuration: Setting up Lambda with Bedrock was time-consuming and complex.
• Game Bugs: Issues like inventory management and NPC behavior testing delayed progress.
• Model Performance: The initial Llama model struggled with repetitive and inconsistent outputs. Switching to the Mistral model greatly improved results.

Accomplishments We’re Proud Of

• Successfully developing our first game using Godot 4.
• Improving Prompt Engineering Skills
• Seamlessly integrating AWS services into gameplay mechanics.
• Gaining hands-on experience with Generative AI tools and AWS technologies.

What We Learned

• Fundamental concepts of Godot 4 and game development workflows.
• Effective setup and use of AWS services like Lambda and Bedrock.
• Strategies to incorporate AI models into game mechanics for enhanced player engagement.

What's Next for Shadows of Tomorrow

We’re excited to continue improving and expanding our game with the following features:

• Fixing existing bugs and polishing gameplay.
• Adding a food system where players can cook meals and discover new recipes.
• Implementing a quest system for enriched storytelling and player engagement.
• Developing a communication system where players repair broken radio cables to connect towns.
• Introducing a dynamic weather system with seasonal variations like summer, winter, and autumn.
• Enhancing the game's UI/UX for a smoother player experience.
• Exploring multiplayer functionality for cooperative gameplay.

Built With

• amazon-bedrock
• amazon-q
• aws-lambda
• godot-4
• mistral-ai