Vision AI product proof Live
BookScout
BookScout turns a bookshelf photo into recognized books, structured results, market context, and personalized reading recommendations.
- 1 photo
- to start a shelf scan
- 3 markets
- represented in price checks
- AI
- vision and recommendation pipeline
Photograph a bookshelf, identify the books, compare prices, and get recommendations.
The product is retained as evidence of a complete AI-native workflow: a defined input, structured processing, a usable output, cloud operations, and real distribution.
- Gemini-powered shelf recognition
- Structured book cards and market prices
- Personalized reading recommendations
- Shelf analysis and playful insight modes
A complete path from user intent to delivered result.
Each step has a clear purpose and observable output.
Identify and normalize visible books
Enrich results with pricing context
Present recommendations and shelf insights
Built as an operating product, not a disconnected demo.
The stack reflects the product's specific needs while staying inside the Factory's GCP, observability, release, and ownership standards.
SYSTEMS USED
- Gemini vision
- Next.js
- Cloud SQL
- Firebase Auth
- Stripe
- Cloud Run