🌱 TouchGrass AI
AI that tells you to stop using AI.
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
TouchGrass AI is a local-AI-powered outdoor mission generator designed to get people away from their screens and physically outside.
Instead of using AI to keep you scrolling, TouchGrass AI uses AI to create a personalized real-world mission based on:
- ⏱️ Available time — 15 min, 30 min, 60 min, or 2 hours
- 🌳 Activity — Walking, Nature, Photography, Fitness, Exploration, or Surprise Me
- 🎯 Difficulty — Easy, Medium, or Challenging
The AI generates a unique outdoor quest with tasks, a suggested distance, difficulty, XP reward, and a final phone-free challenge.
Once the mission starts, the user is encouraged to put their phone away and go outside.
After returning, they can complete the mission, track their XP, streak, and total time spent outdoors.
The project is built for students, developers, remote workers, and basically anyone who spends too much time staring at a screen.
The goal isn't to make AI more addictive. It's to use AI to help you leave the screen.
Demo
🌐 Live Website: https://6069758.github.io/TouchGrassAi/
Try generating a mission, start it, put your phone away, go outside, and come back to complete it.
Code
💻 GitHub Repository: https://github.com/6069758/TouchGrassAi
The entire project is open source and built with vanilla HTML, CSS, and JavaScript.
How I Built It
TouchGrass AI uses local open-weight AI through Ollama.
The current model is:
Qwen3 4B
The architecture is intentionally simple:
USER
│
▼
┌───────────────────┐
│ HTML / CSS / JS │
│ User Settings │
└─────────┬─────────┘
│
▼
┌───────────────────┐
│ Mission Prompt │
│ Builder │
└─────────┬─────────┘
│
▼
┌───────────────────┐
│ Ollama │
│ Local Inference │
│ localhost:11434 │
└─────────┬─────────┘
│
▼
┌───────────────────┐
│ Qwen3 4B │
│ Open-weight AI │
└─────────┬─────────┘
│
▼
┌───────────────────┐
│ Mission JSON │
│ Title / Tasks / │
│ XP / Distance │
└─────────┬─────────┘
│
▼
🌳 OUTSIDE
│
▼
📵 PHONE AWAY
The browser sends the user's preferences to the locally running Ollama instance. Qwen3 generates a structured JSON mission, which the JavaScript application turns into an interactive quest.
The AI isn't just being used to generate some text on the homepage. It is the core engine that creates the actual outdoor experience.
The project also has a fallback demo mission so the interface remains usable if local AI isn't available.
Tech Stack
- HTML
- CSS
- Vanilla JavaScript
- Ollama
- Qwen3 4B
- Local browser storage for XP, streaks, and outdoor time
No paid AI API is required.
Why Does Open Innovation Matter?
Open innovation is especially important for this project because the whole point is to make AI less dependent on the cloud and less focused on keeping people online.
Using local open-weight AI makes several things possible:
🔒 Privacy
User preferences and prompts can remain on their own machine instead of being sent to a third-party AI API.
📴 Local & Potentially Offline
Once the model is downloaded, inference can happen locally without depending on a cloud AI service for every mission.
🧠 Model Freedom
The application isn't locked to a single proprietary model.
Users can experiment with different open-weight models and potentially fine-tune or modify them for different types of outdoor experiences.
💰 No Per-Request AI Cost
There is no paid API call every time someone generates a mission.
This makes the idea much more accessible for students, hobbyists, and developers experimenting with AI.
🌱 AI With a Different Goal
Most AI products try to make you spend more time with the screen.
TouchGrass AI intentionally does the opposite.
The AI's job is to generate something useful and then encourage the user to stop interacting with the AI.
That's the part of open innovation I wanted to explore: not just making AI more powerful, but experimenting with what we actually want AI to help us do.
My Agent Session
The project was built with AI-assisted development, with the application architecture and implementation centered around local open-source AI.
Prize Categories
I am entering:
- 🌱 Touch Grass
- 🤖 Open-Source AI
The core requirement of the project is local open-weight AI, and the resulting experience is specifically designed to encourage users to leave their screens and interact with the real world.
What's Next?
Some ideas I would like to explore next:
- 📸 Local vision AI to verify outdoor mission photos
- 🐦 Bird identification using local vision models
- 🌦️ Weather-aware missions
- 📍 GPS-based exploration challenges
- 🗺️ Offline outdoor routes
- 🏆 Community challenges and leaderboards
- 🌳 Location-specific nature missions
- 🤖 Multiple local models for different types of missions
The bigger idea is simple:
AI doesn't always need to give us another reason to stay online.
Sometimes the best thing an AI can do is give you a reason to close the laptop.
Go touch grass. 🌱
Top comments (0)