ADK Shopping Assistant
A conversational shopping demo that guides visitors from product discovery to a validated cart and order summary. Four Google ADK agents handle the conversation while Python controls catalogue prices and totals; Redis restores each browser’s shopping session.
The problem
A conversational shopping demo that guides visitors from product discovery to a validated cart and order summary. Four Google ADK agents handle the conversation while Python controls catalogue prices and totals; Redis restores each browser’s shopping session.
My contribution
I integrated four ADK agents with deterministic catalogue and cart tools, a streamed browser interface, Redis session recovery, and guarded API requests.
Product features
- Four specialist agents guide profile collection, catalogue browsing, checkout, and order summaries.
- Python validates products and quantities and calculates cart totals from a fixed demo catalogue.
- Streamed responses show progress while the assistant completes each turn.
- Redis-backed browser sessions restore the conversation and cart after reload.
- Uses made-up contact details; no payment is collected or real order placed.
Engineering work
- Agents with deterministic tools: Coordinate four shopping specialists while Python validates catalogue choices, quantities, and totals.
- Durable conversations: Restore an opaque browser session and cart from Redis across requests and reloads.
- Streamed workflow: Return progressive chat events, validate inputs and origins, and keep checkout explicitly simulated.
Outcome & scope
Connects product discovery, cart calculation, and checkout in one conversation. A live synthetic checkout verified 2 × ₹70,000 = ₹1,40,000 and an accurate demo order summary; no sales or conversion results are claimed.
Public shopping simulation · Use made-up details · No real orders or payments