DEV Community

Himanshu Kumar
Himanshu Kumar

Posted on

TouchGrass AI

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

What I Built

I built TouchGrass AI, a nature-inspired outdoor activity companion designed to help people spend less time scrolling and more time experiencing the world around them.

Users choose their available time, mood, energy level, and preferred activity. TouchGrass AI turns those preferences into a personalized outdoor mission, whether it's a mindful walk, a nature observation challenge, or a small photography adventure.

The idea is simple: technology should sometimes help us put technology away.

Instead of keeping users engaged with another endless feed, TouchGrass AI gives them a reason to step outside, complete a real-world activity, and reflect on what they experienced.

Demo

https://touchgrass-ai-n6t4.bolt.host/

Code

https://github.com/Himanshu777685/touchGrassAI.git

How I Built It

I built TouchGrass AI using React, TypeScript, and a lightweight application architecture focused on personalization and a distraction-free user experience.

The application takes a user's preferences and generates structured outdoor missions with clear instructions, a time estimate, and a reflection prompt. It also supports a simple mission-completion and progress-tracking flow.

The AI integration is designed around open-weight language models, with a configurable provider interface and a fallback for situations where the model is unavailable.

Why Does Open Innovation Matter?

For a project like TouchGrass AI, open innovation is about more than just choosing a different AI model.

Open-weight models offer the possibility of running inference locally, keeping personal preferences on the user's device, and building experiences that do not depend on a proprietary inference API. With a suitable local setup, mission generation can even work without an internet connection after the model has been downloaded.

Open approaches also give developers more control over model selection, inference settings, output formats, and application behavior.

I want AI to be a tool that helps people engage with the world—not another reason to stay glued to a screen. Having the freedom to inspect, adapt, and run the underlying technology makes that vision more achievable.

Top comments (0)