Developer creates fully offline nature identification app using Gemma 3 AI
A developer built Trailside, an on-device nature field guide that identifies plants, birds, insects, and fungi through camera photos. The app operates completely offline without cloud services or accounts by running the open-weight Gemma 3 vision model locally via Ollama. It was submitted for the Hacktoberfest Open-Source AI Challenge as a tool designed for remote outdoor use where cellular signals are unavailable. The model performs multimodal identification directly on the user's device without requiring any external AI runtimes or partner services.



