This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.
What I Built
I built TouchGrass — AI Outdoor Quest, an open-source web application that encourages people to spend less time on screens and more time outdoors.
Users choose how much time they have, their preferred activity, difficulty level, and whether they are going alone or with friends. TouchGrass then provides a simple outdoor quest with steps to follow.
The idea is to make going outside easier and more enjoyable, even when someone only has 10 minutes to spare.
Demo
Try TouchGrass: https://touchgrass-k9y7.onrender.com/
The hosted demo uses built-in quests, so anyone can try the experience without installing an AI model. AI inference is not currently running on the hosted deployment.
Code
GitHub repository: https://github.com/nikhilengineer21-bit/TouchGrass
TouchGrass is open source under the MIT License. You can explore the code, report issues, and contribute improvements.
How I Built It
I built the application using HTML, CSS, JavaScript, and Python.
For local AI generation, TouchGrass uses Qwen2.5 0.5B through Ollama. The Python backend sends the user's preferences to the local model, validates its response, and returns a quest when the response passes validation. If AI generation fails or the response is invalid, the app can use a built-in fallback quest.
The hosted version runs on Render and currently uses those fallback quests because the local Ollama model is not running on the server.
Building this project helped me learn how a frontend, backend, local AI model, response validation, and fallback logic work together.
Why Does Open Innovation Matter?
Open innovation gives developers the freedom to experiment, learn, and build useful tools without depending entirely on closed AI services.
Using an openly available model through Ollama allowed me to explore local AI inference without relying on a paid remote AI API. It also makes it possible to experiment with models and understand how they behave in a real application.
When run locally, TouchGrass can generate quests on the user's own computer rather than sending preferences to a third-party AI inference service.
This project is an early prototype, but it shows how open-source software and accessible AI tools can help turn an idea into something people can use, learn from, and improve together.
My Agent Session
I did not record a DevRelay agent session for this project.
What's Next?
I'd like to improve quest variety, strengthen response validation, add automated tests, and make contributing easier for other developers.
My goal is simple: less scrolling, more exploring.
Thanks for checking out TouchGrass! 🌿
Top comments (0)