All work

mDealer

Case 04 of 10

  • SaaS
  • Own product
  • 2026
Visit mdealer.makarov.media

Overview

SaaS for car service shops: booking by service bay, work orders with parts and payments, inventory, a deal pipeline and an API for existing integrations.

Year
2026
Type
SaaS · Own product
Role
Design and full-stack development
478
automated tests
56
data models
25
API modules

Challenge and solution

The challenge

Repair shops juggle bookings, work orders, stock, payments and customer messages across several tools — and their customers get no app at all.

The solution

A multi-tenant SaaS: a Fastify and Prisma API with strict per-shop isolation, a staff app for the whole workflow including 1C export, and a separate app for car owners. A Claude-based assistant sends maintenance reminders and books customers into free slots.

What it does

  • Sign-up for shops with 4 roles and email invites
  • Booking calendar by service bay with double-booking protection
  • Work orders, estimates, diagnostics and invoices with parts and labour
  • Deals on a kanban board with custom stages, plus tasks
  • Inventory: receipts, transfers, write-offs and stock counts
  • “My car” app for owners: history, live repair status, online booking
  • AI assistant that chats with customers over WhatsApp and Telegram

Under the hood

  • Tenant isolation inside Prisma: every query gets the shop’s ID automatically; audit log written by database triggers.
  • Postgres advisory locks stop two bookings from taking the same bay and slot.
  • Claude tool-use loop with a Redis queue that merges a burst of messages into one reply.
  • Export to 1C:Accounting in EnterpriseData XML; chunked uploads with ffmpeg transcoding.
  • Next.js
  • React
  • Tailwind CSS
  • Fastify
  • PostgreSQL
  • Prisma
  • Redis
  • Zod
  • Claude
  • Docker
  • Coolify