Touch Grass Sports AI – Offline AI That Gets You Outside
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass What I Built Touch Grass Sports AI is a fully offline open-source AI that generates personalized outdoor sports challenges. Its only mission: get you off the screen and into the real world. You pick your fitness level, sport, available time, and location. It replies with a concrete challenge that must be done…
This is a Hacktoberfest Open-Source AI Challenge submission named Touch Grass Sports AI. The AI's primary purpose is to motivate users to step away from digital devices and engage in outdoor physical activities. Users provide details such as fitness level, chosen sport, available time, and location. In response, Touch Grass Sports AI offers a specific challenge that must be completed outside, always featuring a "touch grass" element.
The platform accommodates various sports including running, basketball, cycling, yoga, swimming, climbing, parkour, volleyball, tennis, skateboarding, martial arts, hiking, and more.
The AI operates in two modes: Local LLM (Ollama) for generating creative challenges and a pure rule-based mode that does not rely on any model. The demonstration is available through a web-UI and command line interface (CLI). The codebase is open-source and can be found on GitHub at https://github.com/gabutersproject/touch-grass-sports-ai.
The development of Touch Grass Sports AI is centered around open-source AI and local inference. The program employs Ollama for running open-weight models locally, including Llama 3.2, Phi-3, Gemma 2, and others. It enforces strict system prompts to ensure every output promotes outdoor activities and includes a "touch grass" action. Additionally, the AI supports multiple languages, English and Indonesian.
The motivation behind this open innovation is to prevent reliance on closed cloud APIs, which could hinder the original concept. The software functions entirely offline, making it ideal for use in remote locations with limited or no signal. It does not transmit any data from the device, is free of charge, and lacks any API costs.
Users also have the ability to customize the AI by swapping models, modifying prompts, and expanding the range of challenges. The "touch grass" philosophy is at the core of this project, emphasizing a return to real-life experiences and physical activity, rather than relying on dopamine-inducing digital content.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
