GuidePup
AI-powered vision assistant helping visually impaired users understand their surroundings through real-time camera analysis and voice feedback.
🎯 The Problem
Over 2.2 billion people worldwide have vision impairment. Many rely on others to describe their surroundings, read signs, or navigate unfamiliar spaces.
Daily tasks like reading menus, identifying products, or navigating new places require assistance.
Most apps only do OCR or object detection—they don't provide contextual understanding.
Professional assistive devices can cost thousands of dollars.
👥 Users & Impact
Target Users
- Visually impaired individuals seeking daily independence
- Elderly users with declining vision
- Caregivers and accessibility advocates
- Schools and organizations serving blind communities
🎬 Demo
Video Walkthrough
Demo video coming soon
2-3 min walkthrough of core features
To add a video, set the data-video-url attribute with a YouTube, Vimeo, or direct video URL.
Screenshots
🚀 How to Run
Install via TestFlight
- Install TestFlight from the App Store
- Open the invite link (coming soon)
- Tap "Accept" and then "Install"
- Grant camera and microphone permissions when prompted
🏗️ Architecture
┌─────────────────────────────────────────────────────────────┐
│ GuidePup App │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ Camera │───▶│ Frame │───▶│ GPT-4 │ │
│ │ Module │ │ Capture │ │ Vision │ │
│ └──────────────┘ └──────────────┘ └──────────────┘ │
│ │ │
│ ▼ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ Audio │◀───│ Text-to- │◀───│ Response │ │
│ │ Output │ │ Speech │ │ Parser │ │
│ └──────────────┘ └──────────────┘ └──────────────┘ │
│ │
├─────────────────────────────────────────────────────────────┤
│ Storage: AsyncStorage │ API: OpenAI │ Platform: iOS/Expo │
└─────────────────────────────────────────────────────────────┘
Tech Stack
✨ Key Features
Real-time Analysis
Point your camera and get instant AI-powered descriptions of your surroundings.
Voice Feedback
Natural text-to-speech reads descriptions aloud. Adjustable speed and voice.
Context Aware
Understands scenes, not just objects. "A crowded coffee shop with an empty table near the window."
Accessibility First
Built with VoiceOver support, high contrast, and large touch targets.
Offline History
Review past descriptions offline. Useful for remembering locations or items.
Privacy Focused
Images processed in real-time, not stored. Your API key stays on device.
📊 Metrics & Results
Testing Notes
Tested with 5 visually impaired beta users. Key feedback: "Finally an app that tells me what's happening, not just what objects are in frame."
🗺️ Roadmap
Camera capture, GPT-4 Vision integration, voice output
Polish UI, accessibility audit, submit for review
TestFlight rollout, gather user feedback
Step-by-step guidance, obstacle detection
Port to Android, expand reach
🙏 Credits
Charlie Han
Atrak Team
OpenAI GPT-4 Vision API
React Native + Expo
Special thanks to our beta testers from the blind community who provided invaluable feedback.