Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

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

I built TouchGrass AI, an early-stage nature companion designed to encourage people to take breaks from their screens and spend more time outdoors.

The idea is simple: instead of spending more time scrolling, users can discover small outdoor missions that encourage them to explore the real world.

The goal is to make technology a starting point for real-world experiences, rather than another reason to stay indoors.

The project is currently an early prototype. Ready-made missions can be used without connecting an AI model.

A live demo will be added after deployment.

The repository contains the project source code and setup instructions.

The AI integration is designed to use Ollama with the qwen2.5:3b model to generate outdoor missions locally. The AI connection still requires setup and testing.

The project also includes ready-made missions, so users can explore the core idea without configuring the AI component.

Open innovation allows developers to experiment, learn, and build on existing work.

Using an open-weight model with local inference offers an alternative to relying entirely on hosted AI services. Depending on the setup, it can reduce recurring API costs and give users more control over where their prompts are processed.

I want TouchGrass AI to be a small example of how open-source AI can support everyday activities and encourage healthier relationships with technology.

One challenge was making the application useful even when the AI component is unavailable. Ready-made missions help users explore the main idea without depending on a working AI connection.

Building this prototype is also helping me learn about web development, local AI integration, and sharing projects through open source.

Optional: Add a DevRelay session link if available.

This project is submitted for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.