JobiGo — An AI Football Coach That Wants You Off the Screen
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass What I Built Most AI apps want more of your screen time. I built one that wants less. JobiGo is an open-source AI football coach built around one simple idea: The screen should be the shortest part of the experience. It is designed for football players who want a personalized training session without needing a…
JobiGo is an open-source AI football coach that aims to decrease screen time by encouraging physical activity. Unlike many AI applications that want users to spend more time on their screens, JobiGo is designed to be used without needing a coach present. The app's training sessions follow a specific plan: plan, go outside, play, return, report, and improve.
When a user begins, they select their training duration, level, goal, equipment, and playing environment. JobiGo then employs Gemma 3 4B, an open-weight AI model, to create a customized football mission based on the user's input. For the purpose of testing, a 45-minute Intermediate Cone Dribbling Challenge was generated, including a warm-up, cone weave, figure-8 drill, speed dribble, challenge, cooldown, success metric, and safety guidance.
The crucial aspect of JobiGo is to detach the player from the screen once the session starts. The app tells the user to put their phone down and complete the training session. Upon returning, the user reports the actual results: attempts, successful attempts, goals, drills completed, difficulty level, and any notes. Gemma reviews this information and delivers a performance summary, highlights the player's strongest area, identifies an area for improvement, and offers a recommendation for the following session.
This process depends on the previous session's performance rather than just generating a new random workout. For example, in one test session, Gemma identified the player's strengths in dribbling and areas to improve, recommending another dribbling-focused session based on those results.
JobiGo's core is the Gemma 3 4B model, which runs locally using Ollama. The primary architecture consists of a FastAPI backend, frontend, AI provider layer, session storage, validation, and tests. The AI layer is separated into provider abstractions, allowing the incorporation of other inference providers without rewriting the core application.
JobiGo directly addresses the "Touch Grass" theme by employing AI to encourage physical activity while minimizing screen time. The app's core loop is intentionally structured around leaving the screen: PLAN → GO OUTSIDE → PLAY → RETURN → REPORT → IMPROVE.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
