Can I vibe code Sparque?

Build a custom hybrid search and merchandising API service for an e-commerce platform using Node.js/TypeScript, Fastify, and Typesense.

1. DATA ARCHITECTURE
- Design a database schema for Products (id, title, description, category_id, tags, price, in_stock, score_boost).
- Define a collection schema in Typesense for fast hybrid indexing including text fields and numerical attributes for filtering.
- Set up Redis data structures to store real-time user session context (e.g., last viewed categories, clicked tags).

2. CORE FUNCTIONALITY
- Implement a POST /api/search endpoint that receives query string, user_session_id, filters, and pagination params.
- Execute hybrid search querying Typesense for full-text keyword matches while retrieving real-time vector embeddings for semantic matches.
- Implement a post-processing re-ranking layer in TypeScript:
  a. Apply hard business rules: check for pinned products defined in a database table for the specific query.
  b. Apply dynamic boosts: boost items matching the user's recent category affinity (fetched from Redis session).
  c. Apply inventory penalty: drop out-of-stock items by 50% score reduction unless explicitly filtered.
- Expose an admin REST API to manage Merchandising Rules (pin SKU to position N for query string X, boost tag Y by multiplier Z).

3. INTEGRATION & TRACKING
- Implement a POST /api/events endpoint to track clicks, views, and purchases, updating Redis session affinities asynchronously.
- Provide an ultra-fast GraphQL/JSON search API contract for front-end integration (sub-50ms processing target).

4. FAILURE HANDLING & CONSTRAINTS
- Gracefully fall back to plain keyword search if vector calculation or Redis session lookup times out (>15ms).
- Out of scope: Automated deep learning model training pipelines, external multi-tenant SaaS provisioning.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs