OrderFlow — Multi-Tenant Ordering SaaS with an AI Agent

OrderFlow is my own product, not client work — a live multi-tenant SaaS that lets restaurants and food shops take orders over WhatsApp and manage them from a dashboard.

Small food businesses do not want another app their customers have to install. They already take orders on WhatsApp, badly: messages get missed, items get written on paper, and nobody knows what was sold at the end of the day.

What OrderFlow does:

An AI agent handles the conversation — it reads the menu, answers questions, takes the order and confirms it, instead of making the owner type every reply.

Each business gets its own isolated tenant: its own menu, branding, prices and order history, on shared infrastructure.

Orders land in a dashboard the owner can actually use on a phone — accept, mark ready, mark delivered.

The menu supports fixed prices, variants and weight-based pricing, because real shops do not all price the same way.

Order history becomes reporting, so the owner can see what sells without keeping a notebook.

Built as a production system I run and maintain myself: multi-tenancy with row-level isolation, a Next.js front end, a Supabase backend, and monitoring on the paths that matter. Because I own it, every architectural trade-off in it is one I have had to live with afterwards — which is a different kind of experience from shipping and walking away.