Featured projects Get in touch

Food retail · Ordering & shop management · Live storefront

RabahFoods
Ops

Customers order by chat.
Staff record sales by voice.

RabahFoods brings customer ordering and the daily running of a frozen-food shop into one place. An AI assistant helps customers place orders, while staff can speak their sales instead of typing them in.

For
Customers, sales staff & owners
Role
Architecture and full build
Availability
Live storefront
RabahFoods live website showing its frozen-food storefront and Ms Poundz ordering assistant
Public storefrontVisit website ↗

What it makes possible.

01 / What makes it useful

An ordering assistant that takes the next step

Customers tell Ms Poundz what they want on WhatsApp, Instagram or the website. The assistant prepares an order for them to review and confirm, reducing the back-and-forth needed to take an order manually.

02 / What makes it useful

Say the sale. Keep serving.

Staff can dictate completed sales hands-free. Spoken product names, quantities and payments become an editable table, with a review before anything is saved. Less typing, with a chance to catch misheard details.

03 / What makes it useful

One record from order to payment

Confirmed orders, stock changes and payments belong to the same workflow. The team can follow a sale through to completion and see any outstanding balance without piecing together separate records.

The everyday essentials.

From checking what is in stock to knowing who has paid, the everyday details stay connected.

See what is available

Customers browse the frozen-food range on the public shop before starting a conversation with the ordering assistant.

Follow each order

Staff can see which orders are waiting, ready for collection or delivery, and completed.

Keep stock in step with sales

The shop tracks incoming stock and sales by the piece or full box, so the team can see what remains.

Know who still owes

Full and partial payments sit against each sale. Staff can record debt payments and reprint a receipt when a customer needs one.

Make sense of the day’s money

Owners can review sales, expenses and profit, then compare expected cash, mobile money and bank payments with what was received.

Keep customer details together

Orders, balances and payment records stay connected to the customer, making follow-up easier.

For the curiousBehind the buildExplore the architecture and tools.

Architecture

How a conversation, the shop and the back office share one system.

Hover or tap a component to trace its connections.

WhatsApp and Instagram

Where customers message the business

  • WhatsApp Cloud API
  • Instagram Messaging

Meta webhook

Verifies the signature, stores the message, replies 200 at once

  • Next.js route
  • HMAC
  • after()

Ordering agent

Reads the live catalogue, asks questions, drafts the order

  • Tool calling
  • Shared runner

Model provider

OpenAI or Anthropic behind one interface

  • OpenAI
  • Anthropic

Confirmation page

Customer reviews, picks payment and address, confirms

  • Single-use token
  • Places

Firestore

Products, batches, orders, sales, payments, conversations

  • Firestore
  • Admin SDK

Storefront and web chat

The public shop and its chat with the same agent

  • Next.js
  • React

Staff alerts

Pushes each confirmed order to staff

  • Telegram Bot API

Operations app

Stock intake, orders, payments, reconciliation, reports

  • TanStack Table
  • Recharts

Firebase Auth

Admin and sales roles

  • Email and password
All connections
  1. WhatsApp and Instagram to Meta webhook: message
  2. Meta webhook to Ordering agent: agent turn
  3. Storefront and web chat to Ordering agent: /api/chat
  4. Ordering agent to Model provider: prompt + tools
  5. Ordering agent to Firestore: draft order
  6. WhatsApp and Instagram to Confirmation page: confirm link
  7. Confirmation page to Firestore: draft to pending
  8. Confirmation page to Staff alerts: staff alert
  9. Operations app to Firestore: fulfil, collect, report
  10. Operations app to Firebase Auth: roles

Technology stack

What it is built with.

Application

  • Next.js 16 App Router
  • React 19
  • TypeScript

Interface

  • Tailwind CSS 4
  • TanStack Table
  • Recharts
  • React Hook Form

Data

  • Firestore
  • Firebase Auth
  • Firebase Storage
  • Admin SDK

AI and messaging

  • OpenAI
  • Anthropic
  • WhatsApp Cloud API
  • Instagram Messaging
  • Telegram Bot API

Next project

Job search · Personal workspace employed