This is a submission for Weekend Challenge: Dog Days Edition
What I Built
PawID is an AI-powered dog breed identifier built with Google Gemini 3.7 Flash. The idea was simple, every dog owner or dog lover has looked at a dog and wondered "what breed is that?" PawID answers that question instantly.
Upload any dog photo and within seconds the app returns a complete breed profile including:
- Choose Photo from gallery
- Take Photo with camera
- Quick Test with sample dogs (Golden Retriever, Siberian Husky)
- Instant Gemini Vision 2.5 analysis
- Comprehensive Canine Profile results
- Works on mobile and desktop
Screenshots
Demo
👉 https://ai.studio/apps/c3e04c01-ca4c-4e61-af22-912b79e750c8
Try it with any dog photo, the more clearly the dog is visible, the better the results.
Code
👉 https://github.com/elsie-rainee/pawid
How I Built It
Built entirely in Google AI Studio using Gemini 3.7 Flash as the multimodal AI brain.
Core flow:
- User uploads or takes a dog photo
- Image sent to Gemini 3.7 Flash as multimodal input
- Gemini analyzes physical traits like skull structure, coat pattern, body shape, ear type, tail shape
- Returns structured breed data
- App displays results in animated cards with full breed profile
What I Learned
The most interesting challenge was prompt engineering, getting Gemini
to return consistent structured data every single time required careful
design. Gemini's vision model handles mixed breeds impressively, rather
than refusing to answer it identifies dominant breed characteristics, which is actually more useful than a simple classification.
Prize Categories
Best Use of Google AI
Top comments (0)