Inspiration

FORESTER was inspired by the reccuring forest fires that have occured in recent years and the enviromental damage they cause. Some of these fires are linked to illegal land clearing, threatening ecosystems, wildlife, and nearby communities. We wanted to explore this issue through a game that makes forest protection interactive and engaging. By combining tower defense mechanics with AI-powered enemy strategies, we aim to challenge players to think critically about how they protect forest from destructive threats.

What it does

FORESTER is a web-based tower defense game where players must defend a forest against enemies attempting to burn it down. The game consist of two main phases: Preparation and Wave. During the Preparation Phase, players strategically place their defenders while managing a limited number of Player Points (PP). Each unit has different abilities, ranges, and cost, so every placement matters. Once the player start a wave, an AI powered by the Apertus 70B model analyzes the battlefield setup and determines an enemy deployment strategy designed to counter the player's defenses. Players must build effective defenses, anticipate threats, and protect the forest through incoming wave of attacks.

How we built it

FORESTER combines Unity-based gameplay with an AI-powered strategy generation system using the Apertus 70B large language model. The Process begins when the player places defenders across the map using their available Player Points. Once a wave starts, Unity collect the battlefield data, including defender types, positions, attack ranges, and damage values, and sends it to the AI model. The model then analyze the player's setup and generates an enemy deployment strategy. This strategy is returned to the game, where the enemy troops are deployed and the wave begins. This Approach allows enemy encounters to respond to the player's defensive choices rather relying entirely on predefined enemy formations.

Challenges we ran into

One of our main challenges was connecting strategic gameplay with AI-generated enemy deployments. The AI needs to interpret battlefield data and produce a strategy that the game can execute correctly. We also had to consider how to balance resource management, defender placement, and enemy difficulty. The AI's strategy should challenge the player without making the game feel unfair or impossible to win. Another challenge was keeping the gameplay understandable and accessible while introducing an AI system that operates behind the scenes.

Accomplishments that we're proud of

We are proud to bring together environmental awareness, tower defense gameplay, and large language model integration in one project. FORESTER gives players control over their defensive strategy while allowing AI to determine how enemies respond. This creates a different kind of tower defense experience, where success depends not only on placing strong units but also on anticipating how an opponent might counter them. We are also proud to explore how AI can be used as an active part of gameplay rather than simply as a conversational feature.

What we learned

Through FORESTER, we learned how to approach AI integration as part of a gameplay system. We explored how structured battlefield data can be used as input for an LLM and how its generated output can inform enemy deployment. We also learned the importance of balancing AI-generated strategies with clear game rules, predictable data formats, and fair gameplay. Beyond the technical aspects, the project reinforced how interactive experiences can help communicate environmental issues in an engaging way.

What's next for Forester

We plan to improve the AI's ability to generate challenging but balanced enemy strategies, refine the relationship between defender placement and enemy deployment, and continue balancing the game's economy and wave difficulty. We also hope to introduce additional maps, defender types, enemy behaviors, and environmental scenarios inspired by forest conservation. Our long-term goal is to make FORESTER an engaging strategy game that encourages players to think creatively about protecting forests.

Built With

Share this project:

Updates

Submission history