This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
🌿 TouchGrass AI — Less Screen Time, More Green Time
TouchGrass AI is an AI-powered web application designed to encourage people to spend less time on screens and more time exploring the outdoors.
In today's digital world, students, developers, remote workers, and many others spend hours using their devices. TouchGrass AI aims to turn that screen time into an opportunity to discover real-world experiences.
The application suggests outdoor activities based on the user's mood, available time, location, and interests. It also includes outdoor challenges, progress tracking, and a personal activity journal to encourage users to build healthier digital habits.
Key features:
- 🤖 AI-powered outdoor adventure recommendations
- 🌱 Personalized suggestions based on mood and interests
- ⏱️ Activities suited to available free time
- 🎯 Outdoor challenges and missions
- 📈 Progress tracking
- 📔 Personal activity journal
- 🎨 A responsive, nature-inspired user interface
The main idea is simple: instead of using AI to keep people on their screens, use AI to inspire them to step outside.
Demo
🌐 Live Website:
https://shivsi3399.github.io/Touchgrass-ai/
The project is deployed using GitHub Pages.
Code
💻 GitHub Repository:
https://github.com/Shivsi3399/Touchgrass-ai
The repository contains the source code for the website, including the HTML structure, CSS styling, and JavaScript functionality.
How I Built It
I built TouchGrass AI using HTML, CSS, and JavaScript, with an open-weight AI model running locally through Ollama.
Technology stack:
- HTML5: Website structure and content.
- CSS3: Responsive layout and nature-inspired design.
- JavaScript: User interactions, activity generation, challenges, and progress tracking.
- Ollama: Local AI model execution and API access.
- Qwen3 4B: Open-weight language model configured to generate outdoor activity recommendations.
- Local Storage: Saves user progress and journal entries in the browser.
How it works:
- The user enters preferences such as mood, available time, location, and interests.
- JavaScript sends the activity request to the locally running Ollama API.
- The configured Qwen3 model generates a suggested outdoor adventure.
- The website displays the recommendation, including activity details and suggested steps.
- Users can explore challenges and record their outdoor progress.
The basic workflow is:
User Input → JavaScript → Ollama → Qwen3 4B → Outdoor Activity Recommendation
The frontend is hosted on GitHub Pages, while the AI model runs separately through Ollama. Using the AI features requires a compatible local Ollama setup.
Why Does Open Innovation Matter?
Open innovation makes AI more accessible to developers, students, and independent creators.
By using an open-weight model with local inference, I could experiment with AI-powered features without making a paid, proprietary cloud API the core of the application. It allowed me to explore prompt design, API integration, structured AI responses, and the practical process of connecting a language model to a web application.
Local inference can also help keep activity preferences on the user's device when the application is configured to communicate only with the local model. This provides an opportunity to build more privacy-conscious applications.
Most importantly, open tools allow other developers to inspect the code, learn from the implementation, customize the experience, and contribute improvements.
TouchGrass AI demonstrates how open AI technology can be used for something beyond generating text: encouraging people to reconnect with nature and develop a healthier balance between technology and real life.
My Agent Session
I don't have a public DevRelay agent-session link to share yet.
Prize Categories
Open-Source AI
This project explores the use of an open-weight language model through local inference with Ollama.
I'll verify the challenge rules before selecting any additional partner-specific prize categories.
Top comments (0)