Detour: My Open-Source AI Entry for the Touch Grass Challenge (Challenge Week 1: Touch Grass)

This article is a recount of a developer's submission for the Hacktoberfest Open-Source AI Challenge, Week 1: Touch Grass. The submission is for an interactive AI web application named Detour.

Detour is designed to disrupt user's habit of spending time on their screens by offering a refreshing alternative. The app is built around a 3-step questionnaire, which inquires about the user's available time, current mood, and preferred environment. The user's responses are then used to generate a personalized real-world mission, such as outdoor photography, creative writing, or deep self-care reset.

The process of Detour is simple yet effective. Once the mission is assigned, the UI fades away, leaving a distraction-free screen that encourages the user to put their phone away and return when they're ready. This screen is titled 'You're good. Put your phone away. I'll be here when you get back.'

Detour aims to provide a solution to the problem of digital burnout and decision paralysis caused by AI assistants. While AI can perform almost any task, it often encourages users to stay glued to their screens, negatively impacting their physical well-being. Detour flips this script by guiding users through a quick questionnaire and then letting them go offline to engage in an activity that suits their mood and time constraints.

The app is built using React, Vite, and Tailwind CSS, with Framer Motion used for smooth, calming animations. The core logic engine uses open-weight models like Llama 3, accessed via a Node.js/Express backend. This design ensures user privacy, as the application does not rely on a closed ecosystem for generating missions. It also allows for easy customization, as anyone can fork the project and run it locally on their own hardware.

The developer emphasizes that Detour is a step towards finding moments of peace in the real world without relying on a centralized intelligence.