Skip to content

Our repository for the Chalmers AI Society x GoWest 2026 hackathon.

License

Notifications You must be signed in to change notification settings

ekrlstd/cycleeffect

Repository files navigation

Flutter Dart Python FastAPI YOLOv8 On-Device AI

PreVue

Your AI Co-Pilot for Smarter, Safer Driving

On-device AI traffic intelligence • Real-time narration • Privacy-first • AV-ready

Status Platform AI

🎯 The Problem

Every year, 1.35 million people die in traffic accidents worldwide. Most of these are preventable with better situational awareness.

Current navigation apps like Waze and Google Maps rely on crowdsourced data—reactive, delayed, and often inaccurate. They tell you there's traffic ahead, but not why it matters to you.

💡 Our Solution

PreVue is an intelligent traffic assistant that understands your context—your speed, heading, and position—and delivers spoken, actionable insights in real-time.

Waze: "Traffic at intersection"
PreVue: "Three vehicles approaching from your left at the roundabout. Yield before entering."

What Makes Us Different

Feature Waze/Google Maps PreVue
Data Source Crowdsourced (delayed) Government APIs + AI (real-time)
Processing Cloud-based On-device (instant)
Privacy Location tracked Data never leaves device
Context Generic alerts Personalized to YOUR trajectory
Offline Limited Full functionality
Future-Ready Static AV architecture

🚀 How It Works

Real-time pipeline: Trafikverket API → On-device AI → Contextual narration

  1. Fetch real-time traffic data from Trafikverket
  2. Combine with your GPS location, heading, and speed
  3. Process through on-device AI for contextual reasoning
  4. Narrate personalized safety alerts via text-to-speech

📱 App Screenshots

image2 image

Dark mode interface optimized for driving • Interactive map • Real-time narration

✨ Key Features

🗣️ AI-Powered Voice Assistant

Ask questions while driving. Get answers without taking your eyes off the road.

  • "What's the traffic like at the next intersection?"
  • "Are there any accidents ahead?"
  • "How many vehicles at the roundabout?"

🗺️ Smart Intersection Tracking

Automatically calculates and monitors the next 5 intersections based on your route, providing relevant data before you get there.

🔒 Privacy-First Architecture

Your location data never leaves your device. All AI processing happens locally. No cloud. No tracking. No compromise.

Our on-device architecture is also purpose-built for V2V (Vehicle-to-Vehicle) processing. We're not building a traffic app—we're building the communication layer for cooperative autonomous driving of the future.

🏃 Quick Start

Prerequisites

  • Flutter 3.8.1+
  • Python 3.8+
  • Android API 26+ / iOS 12+

Installation

# Clone the repository
git clone https://github.com/ekrlstd/cycleeffect.git
cd cycleeffect

# Install Flutter dependencies
flutter pub get

# Start the backend
cd backend
pip install -r requirements.txt
uvicorn main:app --host 0.0.0.0 --port 8000

# Run the app
cd ..
flutter run

Configuration

Update lib/config.dart with your backend IP:

static const String backendHost = 'YOUR_IP_HERE';
static const int backendPort = 8000;

💼 Business Model

Target Market

  • 500K+ daily commuters in Sweden alone
  • New drivers needing extra guidance
  • Professional drivers (trucking, delivery)
  • Fleet operators requiring safety monitoring

Revenue Streams

Stream Description
Freemium Basic narration free, premium features paid
B2B Licensing Driving schools, fleet management, insurance
V2V Platform Future licensing for autonomous vehicles

👥 Team

Built with ❤️ by a team passionate about software, AI, and building things that help people. Aakrish Lama

Aakrish Lama • Abdirashiid Sammatar • Ebbe Karlstad • Syuash Mullick

📄 License

GNU General Public License v3.0 — See LICENSE for details.

PreVue — Because every second counts on the road.

Built for Safer Roads

About

Our repository for the Chalmers AI Society x GoWest 2026 hackathon.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •