Earth Detective: AI That Turns the Real World Into Your Investigation

This submission aims to bring the world outside into an interactive AI-powered game called Earth Detective, which encourages users to disconnect from their screens and engage in real-world investigations. The game offers various themes, such as nature, water, urban environments, and sustainability, to guide players in solving environmental mysteries.

Players receive a mystery, observe their surroundings, record findings, and solve the case. The game includes missions, clues, hints, a timer, difficulty-based scoring, investigation history, and a final field report. Earth Detective is designed for students, explorers, and anyone who wants a meaningful reason to venture outside and explore the world around them.

The project is open-source, built using HTML5, CSS3, Vanilla JavaScript, LocalStorage, and Fetch API, and utilizes Ollama for local AI inference. The AI can generate missions, create environmental clues, provide hints, evaluate observations, and generate a final report. Open innovation is crucial for Earth Detective, as local inference enables the AI to run on the user's machine, promoting privacy, experimentation, and easier modification and extension.

This approach allows developers to explore different open-weight models and create new investigation mechanics without redesigning the entire application around a proprietary API, making the project more adaptable and suitable for its goal of fostering interaction between users and the physical world.