What if AI didn't try to keep you on your screen—but encouraged you to put your phone away and explore the real world?
That's the idea behind NatureQuest AI, my project for the Hacktoberfest 2026 Open-Source AI Challenge: Touch Grass.
🌱 What I Built
NatureQuest AI is an AI-powered outdoor mission coach that turns a few simple preferences into a short, personalized outdoor quest.
Users choose:
⏱️ How much time they have
🌳 Their environment, such as a park, garden, or campus
💭 Their mood
⚡ Their energy level
🎯 Their preferred difficulty
The AI generates one outdoor mission designed around these choices. Once the user completes it, they return to describe what they actually noticed. NatureQuest AI then creates a short reflection grounded in their observations.
The goal is simple: spend less time interacting with AI and more time experiencing the world around you.
🎬 Demo
Watch NatureQuest AI in action:
The demo shows the complete experience: creating a quest, putting the phone away, completing the activity, and reflecting on real-world observations.
💻 Code
GitHub repository:
https://github.com/ChikkalaSukrutiNaidu/NatureQuest-AI
The source code is publicly available so others can explore, learn from, and build on the project.
🛠️ How I Built It
I built NatureQuest AI using:
Python — application logic and AI integration
Streamlit — interactive user interface
Groq API — access to a hosted open-weight language model
OpenAI GPT-OSS 20B — the open-weight model used to generate outdoor quests and reflections
python-dotenv — local environment-variable configuration
The application follows a simple loop:
The user shares their preferences.
The AI generates one outdoor mission.
The user puts their phone away and goes outside.
After completing the mission, the user describes their observations.
The AI creates a reflection based on the user's description.
I also designed the prompts to discourage activities requiring special equipment and to avoid inventing details in the user's reflection.
🌍 Why Does Open Innovation Matter?
Open innovation makes AI experimentation more accessible to students and independent developers like me.
Using an open-weight model through an accessible inference API allowed me to explore how generative AI can support an idea beyond a conventional chatbot. Instead of building an application designed to maximize screen time, I wanted to use AI to encourage people to step away from their devices.
The project also demonstrates how a relatively simple combination of Python, an open-weight model, and an interactive interface can become a practical experiment in human-centered AI.
I hope others can build on this idea with new quest types, accessibility improvements, additional environments, and better ways to help people reconnect with the world around them.
🌿 My Agent Session
I used AI-assisted development to help implement the application, refine prompts, and troubleshoot integration issues. I tested the quest-generation and reflection flows before recording the demo.
One important lesson was that the AI's output matters as much as its interface: quests should be safe and realistic, and reflections should stay grounded in what users actually report.
🏆 Prize Categories
This project was created for the Hacktoberfest 2026 Open-Source AI Challenge — Week 1, “Touch Grass.”
It focuses on using open-weight AI to encourage real-world outdoor experiences rather than prolonged screen interaction.
Final Thought
AI doesn't always need to give us another reason to stay online.
Sometimes, its best job is to give us a reason to step outside. 🌱
Top comments (0)