DEV Community

Viraj Hudlikar
Viraj Hudlikar

Posted on AI-assisted

TouchGrass Buddy: An Offline AI-Powered Outdoor Adventure Generator 🌿

#devchallenge #hf26challenge #opensource #ai

Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass Submission 🌿

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

Most AI applications today are designed to keep users engaged with screens for as long as possible.

For the Touch Grass challenge, I wanted to explore a different idea:

What if AI could encourage people to spend less time on screens and more time outdoors?

That idea became TouchGrass Buddy, an offline AI-powered outdoor adventure generator built using Streamlit, Ollama, and Llama 3.2.

The application generates personalized outdoor missions based on:

  • Available Time
  • Location
  • Adventure Type

Users can choose adventure styles such as:

  • 🌳 Nature Exploration
  • 📷 Photography
  • 🏃 Fitness
  • 👨‍👩‍👧‍👦 Family Fun
  • 🧘 Mindfulness

The AI then creates a personalized outdoor mission designed to get users exploring the real world instead of spending more time online.

Example mission:

🌿 Neighborhood Mindfulness Walk

1. Observe three different sounds around you
2. Find something with an interesting texture
3. Spend two minutes focusing on your surroundings
4. Notice a plant you've never paid attention to before
5. Walk a route you've never taken
Enter fullscreen mode Exit fullscreen mode

The app also includes:

  • Mission generation
  • Progress tracking
  • Mission completion tracking
  • Start New Adventure workflow

The goal is simple:

Generate Mission
       ↓
Go Outside
       ↓
Complete Challenges
       ↓
Track Progress
       ↓
Start New Adventure
Enter fullscreen mode Exit fullscreen mode

The screen should be the shortest part of the experience.

Demo

Home Screen

Home Screen

Generated Mission

Generated Mission

Mission Complete

Mission Complete

Code

GitHub Repository:

👉 https://github.com/vhudlikar/touchgrass-buddy

How I Built It

TouchGrass Buddy is built entirely using open-source tools:

  • Python
  • Streamlit
  • Ollama
  • Llama 3.2

The workflow is straightforward:

  1. The user selects:

    • Available Time
    • Location
    • Adventure Type

  2. The application sends these inputs to a locally running Llama 3.2 model through Ollama.

  3. The model generates a customized outdoor mission.

  4. The user tracks progress using the built-in mission tracker.

  5. Once completed, the user can start a brand-new adventure.

The entire application runs locally on a laptop without requiring internet connectivity after setup.

Why Does Open Innovation Matter?

Open-source AI was not just a feature of this project. It was the foundation of the entire solution.

By using Ollama and Llama 3.2, TouchGrass Buddy can run completely offline while still providing personalized experiences.

This open-source approach provides several benefits:

✅ No internet connection required

✅ No API keys

✅ No recurring costs

✅ No vendor lock-in

✅ User data never leaves the device

✅ Models can be swapped or customized

A closed AI service would have required cloud access, API management, and ongoing usage costs.

For this project, the ability to run AI entirely on a local machine aligned perfectly with the goal of creating a private, portable, and low-cost outdoor adventure generator.

Open innovation made it possible to experiment quickly, build freely, and keep the experience accessible to anyone with a laptop.

My Agent Session

I did not record a DevRelay session for this project.

Instead, I documented the development process through Git commits, project screenshots, and the public GitHub repository.

Prize Categories

🏆 Best Open-Source AI Project

TouchGrass Buddy uses open-weight AI as a core part of the application through:

  • Ollama
  • Llama 3.2
  • Local inference

The project demonstrates how open-source AI can power useful real-world experiences while remaining completely offline and privacy-friendly.

🏆 Best Use of GitHub Copilot

GitHub Copilot assisted with:

  • Streamlit development
  • Python code generation
  • Debugging
  • UI refinement
  • Development acceleration

Copilot helped speed up implementation while allowing me to focus on the overall user experience and challenge goals.

What I Learned

This project reinforced an interesting idea:

AI doesn't always need to increase screen time.

Open-source AI can be used to support real-world experiences rather than replace them.

Building TouchGrass Buddy also helped me deepen my understanding of:

  • Local AI workflows
  • Ollama
  • Llama 3.2
  • Streamlit application development
  • Prompt engineering
  • Privacy-focused AI design

Future Enhancements

Potential future improvements include:

  • GPS-aware adventures
  • Mobile-friendly experience
  • Nature scavenger hunts
  • Achievement badges
  • Group adventures
  • Local park recommendations
  • Wellness and fitness integrations

Conclusion

TouchGrass Buddy started with a simple question:

Can AI help people spend more time outdoors instead of keeping them online?

Using Streamlit, Ollama, and Llama 3.2, I built an offline AI-powered outdoor adventure generator that encourages exploration, movement, mindfulness, and real-world experiences.

The project demonstrates how open-source AI can be used not only to create intelligent applications, but also to encourage users to step away from the screen and engage with the world around them.

GitHub Repository:

👉 https://github.com/vhudlikar/touchgrass-buddy

Top comments (0)