nestjs-multichannel-ai-chatbot-engine

Reusable NestJS AI Chatbot Backend is a reusable backend for catalog and ordering businesses. It combines OpenAI tool calling, pgvector RAG, structured catalog data, persistent memory and deterministic orders. Web HTTP and WhatsApp Cloud API adapters share the same conversational core.

The system searches products, promotions and business knowledge through PostgreSQL for structured data and pgvector for semantic queries. It persists conversations and handles message retries idempotently. Orders are managed with application-controlled prices, totals, availability and transitions. OpenAI interprets language and selects typed tools, while PostgreSQL owns business facts, orders and memory; the model does not determine prices, totals or state transitions.

The scope is one deployment serving one business and database; it is not a multi-tenant SaaS. The current domain supports configurable catalog and ordering, but does not include payments, kitchen or delivery operations, real-time inventory, user authentication or an administrative panel. WhatsApp integration processes messages in the background after returning 200 to Meta, and is a single-process integration without a durable queue or recovery worker.