GrassToucher: Bridging Open-Source AI and Nature for Hacktoberfest 2026

Описание и применение

GrassToucher – это веб‑приложение, использующее открытые модели AI для генерации контекстно‑ориентированных задач на свежем воздухе, помогая разработчикам уменьшить время, проведённое перед экраном.

Задача

Developers spend long hours indoors, leading to screen fatigue and reduced outdoor activity.

Результат

GrassToucher delivers instant, privacy‑preserving outdoor prompts, reducing screen fatigue and encouraging regular nature breaks.

Рабочий процесс

• 1. User opens the web app and grants geolocation permission.
• 2. The app captures latitude, longitude, daylight hours, and seasonal context.
• 3. A lightweight open‑weight model (Google Gemma 2) processes the data and outputs a JSON‑structured outdoor micro‑task.
• 4. The app displays the challenge and prompts the user to lock the screen and step outside.
• 5. User completes the task and returns to the app for the next prompt.

Доказательства

GrassToucher acts as an intelligent environmental companion that evaluates real-world parameters (location coordinates, time of day, weather, ambient sensory context) and provides immediate, context-aware outdoor micro-challenges and ecological observations.

Источник

The application captures client-side environmental metrics (latitude, longitude, ambient daylight hours, and seasonal context) and formats them into a structured system prompt.

Источник

Instead of relying on monolithic closed APIs, GrassToucher utilizes compact open-weight models fine-tuned to parse structured JSON outputs.

Источник