This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
SunStride — A screen-free outdoor companion powered by local Gemma 3 1B AI that generates personalized audio briefings in 10 seconds, then encourages you to put your phone away and enjoy nature completely offline.
Problem: Outdoor apps keep you staring at screens checking weather, UV charts, and fitness widgets.
Solution: SunStride takes 10 seconds of input, generates AI-powered guidance, and literally tells you to close the app and go outside screen-free.
Who it's for: Nature lovers, runners, hikers, privacy advocates, and anyone fighting phone addiction.
Demo
🎥 Live App: https://sunstride.onrender.com/
Try it in 90 seconds:
- Set: Duration (5-60 mins), Activity (Walk/Run/Garden), UV Level
- Click: "🚀 Prepare My Session & Lock Screen"
- Listen: AI-generated audio briefing (~30 sec)
- Go: Put phone in pocket, enjoy nature screen-free
Example output:
"You're heading out for 20 minutes. The sun's moderate—grab water and reapply sunscreen in 15 minutes. Notice the tree canopies. Phone in your pocket now. Enjoy!"
Code
AslamSujah
/
sunstride
Screen-Free Outdoor Sunlight & UV Companion | Powered by Local Gemma & Ollama
☀️ SunStride: Screen-Free Outdoor Companion
Submitted for Hacktoberfest Open-Source AI Challenge - Week 1: Touch Grass
SunStride is an offline-first, screen-free outdoor session planner powered by open-weight local AI (Google Gemma via Ollama). Spend less time on screens and more time touching grass.
🌿 The Problem We Solve
Modern health tools keep us glued to vibrant screens:
- Checking live weather dashboards before outdoor activities
- Monitoring UV charts while walking
- Staring at fitness widgets during park visits
SunStride flips the script: You spend 10 seconds preparing on your laptop, then step outside screen-free with audio guidance.
🎯 How It Works
- Open SunStride → Set your activity duration, type, and current UV level (10 seconds)
- Generate Briefing → Local Gemma model creates a personalized audio script
- Hit Play → Listen to the 30-second safety & mindfulness tip
- Put Phone Away → Entire outdoor session with zero screen time
- Enjoy → Audio reminders…
Repository: github.com/AslamSujah/sunstride
License: MIT | Status: Production-ready on Render
Tech Stack:
- AI: Google Gemma 3 1B (256 MB, <2s inference)
- Inference: Ollama (100% local)
- UI: Streamlit (minimal, distraction-free)
- Backend: Python 3 + SQLite
- Audio: espeak-ng / say / pyttsx3 (cross-platform offline TTS)
- Deploy: Docker + Render
4 dependencies only: streamlit, ollama, pyttsx3, python-dotenv
How I Built It
Architecture
User Input (10s) → Streamlit UI → Ollama Server (Local) → Gemma 3 1B → Text-to-Speech → Audio File → Browser Player → User Goes Outside (100% Offline) → SQLite Journal (Local Logging)
Code
Why Gemma 3 1B?
| Aspect | Gemma 3 1B | Llama 7B | Closed API |
|---|---|---|---|
| Size | 256 MB ✅ | 4 GB ❌ | Cloud |
| Speed | <2s ✅ | 8-15s ❌ | 1-2s |
| Privacy | 100% Local ✅ | 100% Local ✅ | No ❌ |
| Cost | $0 ✅ | $0 ✅ | 💰💰💰 |
| Safety | ⭐⭐⭐⭐⭐ | ⭐⭐⭐ | ⭐⭐⭐ |
Gemma won because it's small, fast, safe, and runs locally.
Top comments (0)