Show HN: Open-Source WhatsApp Sales Agent (Node.js, SQLite, OpenAI Calling) (github.com)

A new open-source AI Sales Agent for WhatsApp has been launched, enabling users to place orders directly through natural language conversations. This innovative tool leverages large language models (LLMs) to convert phrases like "I want 2 milks and an apple delivered to Downtown Dubai" into structured SQLite orders, streamlining the ordering process by eliminating the need for traditional menu navigation. Customers can enjoy a frictionless ordering experience, and businesses can benefit from improved conversion rates in the competitive WhatsApp Commerce space. The significance of this project lies in its combination of AI-driven intent recognition and user-friendly engagement, offering a promising enhancement to the existing landscape of WhatsApp automation, which has largely been dominated by simple bots. Technically, the system is built on Node.js and SQLite, incorporates a hybrid architecture that allows for both AI-processed and traditional flows, and provides privacy options through local LLMs. With features like persistent user sessions, automated cart recovery nudges, and an easy setup, this solution is poised to attract developers looking to implement AI-rich customer interactions in a fast-evolving e-commerce environment.