The scenario
ThreadShop pairs a normal product catalogue with a typed AI stylist. Natural-language requests become structured filters, the assistant assembles outfits from items that are actually in stock, and it explains its picks. Pydantic models keep every tool input and output validated.
System architecture
01 Shopper Describes the occasion
02 Stylist agent Filters + outfit logic
03 Catalogue In-stock items
04 Outfit cards With reasons
High-level flow of ThreadShop, from input to output.
What you will learn
- Natural language to structured product filters
- Typed tools and validated outputs
- Recommendation explanations
What you will build
- Catalogue API
- Stylist agent
- Storefront with chat
Build plan
- Catalogue and API
- Filter extraction
- Outfit builder
- Storefront
- Deploy