Alex Parts

About the project

Alex Parts is a product built for repair shops and spare-parts sellers. It does not focus only on finding the right part. The goal is to unify the whole workflow around that part: vehicle identification, compatibility checks, price comparison, ordering, customer records, stock context and follow-up actions.

In practice, that means replacing constant switching between catalogs, e-commerce systems, CRM tools, WhatsApp and internal admin screens with one conversational layer over the whole operation.

What the product solves

A typical repair-shop workflow is full of small steps that are individually simple but operationally expensive:

• find the correct part,
• verify compatibility through VIN or OEM data,
• compare prices and availability,
• add the item to cart or create an order,
• connect the action to the customer and vehicle,
• continue with follow-up, invoicing or scheduling.

Alex Parts is designed so that one AI agent can coordinate those steps through a dedicated toolset, persistent memory and integrations across the rest of the system.

Architecture

The application runs as a single Node.js process handling both HTTP and WebSocket communication in the same runtime. That matters for realtime chat, agent-run synchronization and admin notifications.

The main layers are:

• Next.js application for the UI and API route layer
• custom Node server for the shared web + websocket runtime
• AI runtime with tool registry and execution layer
• Prisma + MySQL as the data layer for conversations, garage CRM, inventory, invoices and operational records
• WhatsApp integration as a second interface next to the web chat

A particularly strong detail is reconnect handling: if the client disconnects during an agent run, the run continues on the server and syncs back when the client reconnects.

Tech stack

FRONTEND   → Next.js 16 + React 19 + Tailwind
BACKEND    → Next.js App Router + custom Node server
REALTIME   → WebSocket (ws)
DATA       → MySQL + Prisma
AUTH       → NextAuth v5
AI         → AI SDK + OpenAI-compatible provider
MEMORY     → Mem0 long-term memory
CHANNELS   → web chat + WhatsApp Cloud API
I18N       → cs / en / uk / ru

According to the project documentation and implementation, the system uses 26 specialized AI tools for vehicle lookup, VIN decoding, cart actions, invoicing, calendar handling, inventory workflows, escalations and other operational tasks.

Key capabilities

• Natural language across the full workflow — users do not operate isolated forms only; they can describe the request conversationally
• VIN and catalog logic — the system works with VIN decoding, OEM data and catalog sources
• Realtime chat — the web interface streams status, parts, diagrams and agent progress live
• WhatsApp as a real operating channel — webhook handling, batching, follow-ups and connected actions
• Garage / CRM / inventory / invoicing — one data model spans customers, vehicles, service history, stock movements and documents
• Memory and escalation — the agent keeps context and can hand work off when needed

Who it is for

Alex Parts is especially relevant for:

• repair shops that need to speed up part-related work and customer operations,
• spare-parts sellers handling a high volume of requests and orders,
• B2B workflows where speed, compatibility accuracy and reduced manual switching matter.

So this is not just a chatbot for end users. It is an operational and commercial tool for people who deal with concrete service tasks every day.

Project value

The strongest part of Alex Parts is integration depth. Instead of an AI demo on top of a catalog, it becomes a system that can continue into the next real step: cart, order, calendar, garage record, follow-up or escalation.

From an engineering perspective, it is also interesting because it combines an agent runtime, realtime layer, multiple communication channels and a broad operational data model in one product. That makes it a real product-engineering project, not just an AI experiment.

Outcome

Alex Parts is a strong example of how AI becomes useful in a concrete vertical use case — not as a generic chatbot, but as a tool that reduces switching between systems and helps carry work from question to action.

For a portfolio, it showcases AI product design, backend architecture, realtime communication, integrations and operational thinking all at once.