This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
Modern life keeps developers, students, and remote workers in front of screens for hours every day. Even when we have 15–30 minutes of free time, we often end up scrolling through social media instead of taking a meaningful break.
I built TouchGrass AI, an AI-powered outdoor micro-mission generator designed to help people break the doomscrolling cycle and reconnect with the physical world.
Instead of generic advice like "take a walk," TouchGrass AI generates personalized outdoor challenges based on available time, activity preference, energy level, and environment.
Key features:
- 🌿 Sensory-based missions: Explore your surroundings by noticing different shades of green, listening for distinct sounds, and observing nature.
- 📵 Phone-away rule: Encourages users to put their devices away and minimize distractions during their missions.
- ⏱️ Interactive timer and checklist: Track mission progress without constantly checking your phone.
- 🧠 Post-mission reflection: Reflect on breathing, visual fatigue, and mental clarity before and after taking a break.
TouchGrass AI is designed for developers, students, remote workers, and anyone who wants to make short breaks more intentional.
Demo
🌐 Live Web Application: https://hacktoberfest2026-week1.vercel.app/
Try selecting a short walking mission or use the sample mission option to explore the experience.
Code
💻 GitHub Repository: https://github.com/sujitkr268/Hacktoberfest2026_week1
The project is open-source under the MIT license, allowing developers to explore the codebase and build on the idea.
How I Built It
Tech Stack
- Framework: Next.js 16 with App Router
- Frontend: React 19 and TypeScript
- Styling: Tailwind CSS v4
- Icons: Lucide React
- Deployment: Vercel
Open-Weight AI Integration
TouchGrass AI uses an OpenAI-compatible API interface to integrate open-weight language models. The intended options include GPT-OSS 120B and Qwen3.8-27B through Groq, alongside local inference using Ollama.
The application handles AI requests through server-side Next.js Route Handlers, keeping API credentials out of client-side code.
Structured JSON responses allow the application to process generated mission content, while local storage supports mission history and user statistics without requiring an account.
The project explores how a flexible AI integration can support both cloud-based inference and local model execution.
Why Does Open Innovation Matter?
An application designed to help people disconnect from digital distractions should give developers and users greater control over the technology behind it.
Open-weight AI makes this approach possible in several ways:
- Flexibility: Developers can adapt model choices and inference providers to their needs.
- Reduced vendor lock-in: An OpenAI-compatible interface can simplify switching between supported providers.
- Local inference: Ollama allows supported models to run locally after downloading them, reducing dependence on cloud inference for generation.
- Transparency and collaboration: An open-source codebase lets other developers inspect the implementation, learn from it, and contribute improvements.
This project gave me an opportunity to explore how open-weight AI can address an everyday problem rather than creating another reason to stay glued to a screen.
Sometimes, useful technology is technology that helps us step away from it. 🌱
My Agent Session
No public DevRelay agent session is included in this submission.
Prize Categories
Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
Top comments (0)