This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
🌿 Touch Grass AI — Your Offline Outdoor Companion
Touch Grass AI is an AI-powered outdoor companion designed to help people spend less time staring at screens and more time exploring the real world.
The idea is simple: tell the app how you feel, how much time you have, and what kind of outdoor experience you want. It suggests simple activities such as taking a nature walk, exploring a nearby park, observing birds, starting a small garden, or taking a break from technology.
Instead of encouraging people to keep scrolling, Touch Grass AI turns AI suggestions into real-world actions.
It's designed for students, remote workers, and anyone who needs a little encouragement to step outside and reconnect with nature.
Demo
A live demo or video walkthrough will be added here once the application is ready.
Code
The source code will be available on GitHub.
GitHub Repository: [Add your repository link here]
How I Built It
Touch Grass AI is designed around open-weight AI and local inference.
The planned tech stack includes:
- Python for the application logic.
- Ollama to run language models locally.
- Qwen, an open-weight language model, to generate personalized outdoor activity suggestions.
- Streamlit to build a simple, user-friendly interface.
Users provide basic information such as their available time, preferred activities, and energy level. The local AI model uses this information to suggest practical activities that encourage them to go outside.
The application can be designed to work offline after the model has been downloaded, making it useful even when an internet connection is unavailable.
The focus is on keeping the interaction short: get a suggestion, put the phone away, and go outside.
Why Does Open Innovation Matter?
Open innovation makes Touch Grass AI more flexible, private, and accessible.
With local inference, personal preferences and conversations can stay on the user's device instead of being sent to an external AI provider for every request.
Open-weight models also give developers the freedom to experiment with different models, customize prompts, and adapt the assistant to different users and environments.
Unlike an application that depends entirely on a paid, closed AI API, this approach can avoid per-request API fees and potentially work without an internet connection.
Most importantly, the technology supports the project's purpose. AI should not always demand more of our attention. Sometimes, its best job is to give us a reason to put our devices down and experience the world around us.
My Agent Session
My DevRelay agent session will be added here if available.
Prize Categories
To be determined based on the eligible partner categories and the tools used in the final implementation.
Top comments (0)