iGarage
Building an automotive experience around the entire ownership journey.
- ROLE
- Product Experience · Frontend Development · Responsive UI · Commerce Flows
- PRODUCT TYPE
- Multi-vendor automotive marketplace — storefront + vendor panel
- STOREFRONT
- Next.js 15 · TypeScript · Tailwind CSS · MedusaJS 2 · Algolia · Stripe · Strapi · n8n
- VENDOR PANEL
- React + Vite · TanStack Query / Table · Medusa UI · Stripe Connect
- FOCUS
- Discovery · Buying & selling · My Garage · AI assistance · Vendor operations
AT A GLANCE
2
Apps
Shopper storefront + vendor panel
40+
Storefront routes
Localized App Router pages
35+
Vendor panel areas
Orders, inventory, promotions, payouts
4
Design specs
Written before build, with HTML prototypes
CONTEXT
Automotive ownership is a journey, not a transaction
The automotive space is fragmented. Buying, servicing, selling, and discovering vehicles all happen across disconnected platforms. Owners jump between marketplaces, dealer sites, and booking tools with no continuity.
iGarage was built to change that — a platform designed around the complete ownership journey: discovering vehicles, understanding service history, booking maintenance, and buying or selling cars, all in one place.
The challenge: make this feel like a consumer product, not an automotive database. People buying cars are making major decisions. The interface needs to earn their trust.
Discovery
Algolia-powered search, categories, and collections. Compare up to four vehicles side by side, like-for-like.
Buying
Product detail, test-drive scheduling, offers, and Stripe checkout with order confirmation and returns.
Selling
Sell-your-car valuation flow and seller storefronts with public reviews.
My Garage
Per-user vehicles with a garage quota, drivers, and the vehicle as the hub for service and history.
AI Assist
iGarageAI, an n8n-powered chat assistant, plus DiagnostiQ AI vehicle-diagnosis requests with per-user quotas.
EV & Hybrid
Dedicated go-electric and go-hybrid hubs with EV tools for buyers weighing a switch.
Vendors
Vendor panel: products, inventory, orders, reservations, promotions, price lists, and Stripe Connect payouts.
Messaging & i18n
Buyer–vendor messaging (TalkJS), organization invites, and multi-language routing with next-intl.
SERVICE REMINDERS
- Brake padsDue
- Oil changeDue
- TyresOK
Pick a service to book.
Services are listed against the owner’s own vehicle — prices and time up front. (Pick one.)
Owners describe symptoms in plain language — no need to know what part is failing.
Fragmented ownership experience
Owners had no single place to track their vehicle — service history, upcoming maintenance, and buying/selling all lived in separate apps or spreadsheets.
Buying with no guidance
Vehicle listings without intelligent assistance mean buyers had to do all the research themselves. Comparing specs and understanding value required leaving the platform.
Service booking friction
Booking vehicle service typically meant phone calls or third-party booking tools. No context from the vehicle history meant repeating information every time.
MY ROLE
Product, design, and frontend across the full platform
Product Experience
Designed the overall product flow — how users move between discovery, service, and commerce without friction.
Frontend Development
Built the Next.js/TypeScript/Tailwind frontend. Responsible for components, layouts, and responsive behavior across all surfaces.
Responsive UI
Designed and implemented mobile-first responsive layouts. Automotive browsing happens heavily on mobile — the experience had to work there first.
Commerce Integration
Integrated Medusa for product/order management. Designed and implemented the buying and selling flows from browse through checkout.
AI Chat Integration
Connected n8n-powered AI assistant to the platform. Designed the chat interface and the integration flow between user intent and relevant vehicle/service data.
CI/CD
Set up and maintained deployment pipeline. Automated builds and deployment to ensure reliable releases.
Vehicle as the central object
Rather than organizing around actions (buy, sell, service), the interface organizes around the vehicle itself. A user's vehicle becomes the hub — its history, upcoming service, and market value all accessible from one profile.
Mobile-first for discovery
Browsing vehicles is a mobile behavior. The discovery and listing views were designed mobile-first, with the desktop experience as a wider version of the same layout — not a different product.
AI as a navigation aid, not a gimmick
The n8n-powered AI chat was integrated to help users who were uncertain — about which vehicle fit their needs, what service was due, or how to compare options. It surfaces relevant content rather than replacing the browsing experience.
Commerce through Medusa, not custom-built
Building a commerce backend from scratch was out of scope. Using Medusa provided a solid, extensible commerce layer — allowing focus on the user experience rather than payment plumbing.
Decided on paper, proven in a prototype, then built
Significant changes start as written specs with options and a recorded decision, get prototyped in plain HTML, and only then move into React.
Written design decisions
- Products listing: card style and hybrid filter layout chosen from competing options
- Car compare: a verdict layer, like-for-like comparisons only, no invented scores
- Go-hybrid page and My Garage homepage section
HTML before React
- My Garage prototype
- Fuel-log and vehicle-detail sidebars
- Mobile bottom-sheet filters and skeleton states
Brand guidelines & tokens
- Color, type, spacing, radius, elevation, motion
- Dark mode and accessibility standards
- Design tokens JSON + Storybook component docs
Next.js 15 / TypeScript
App Router with localized routes (next-intl), server components for catalog pages, client components where interactivity demands it.
Tailwind CSS + tokens
Design tokens and brand guidelines as the source of truth; Storybook documents the components; full dark mode.
MedusaJS 2
Multi-vendor commerce engine for products, carts, orders, and returns — shared by the storefront and the vendor panel.
Algolia + Strapi
Instant search and filtering over the catalog; Strapi 5 CMS for editorial content like drivers, stickers, and vehicles.
Stripe + TalkJS
Checkout and Stripe Connect vendor payouts; real-time buyer–vendor messaging.
n8n AI integration
n8n workflows power the chat assistant and AI vehicle-diagnosis requests, with configurable per-user quotas.
React + Vite vendor panel
TanStack Query and Table for data-heavy operations screens; drag-and-drop ordering and analytics charts.
Delivery
Standalone Docker builds, GitLab CI pipelines, and Sentry error monitoring.
“Automotive is one of the heaviest purchase decisions a person makes. The interface has to earn that trust.”
Every friction point in the browsing or buying flow risks losing someone at the worst moment — right when they were ready to commit.
Behind the interface
Every screen above sits on a working system. This is the path from a shopper's tap to the services that answer it — which layers I designed and built, and which I configured and integrated.
01 EXPERIENCE
- Design specs
- HTML prototypes
- Tokens + Storybook
Decisions written and prototyped before they reach React.
02 FRONTEND
- Next.js 15 storefront
- React + Vite vendor panel
Localized App Router routes; TanStack Query and Table for vendor operations.
03 API
- MedusaJS 2
- Strapi 5 CMS
Products, carts, orders, and returns for both apps; editorial content from Strapi.
04 DATA
- PostgreSQL
- Algolia
Commerce data in PostgreSQL behind Medusa; Algolia for instant search and filters.
05 INTEGRATIONS
- Stripe + Connect
- n8n workflows
- TalkJS
Checkout and vendor payouts; AI chat and diagnosis with per-user quotas; buyer–vendor messaging.
06 DELIVERY
- Docker
- GitLab CI
- Sentry
Standalone builds, CI pipelines, and error monitoring.
ONE REQUEST, END TO END
- Owner describes a symptom in chat
- Next.js chat UI
- n8n workflow
- Vehicle profile + quota check
- Reply with a bookable service
Core Discovery
Vehicle browsing and listing. The fundamental product — find and compare vehicles with clarity.
Service Layer
Service history and booking added. Vehicle profiles enriched with maintenance context.
Commerce
Medusa integration. Full buy/sell capability with checkout and order management.
AI Integration
n8n AI chat connected. Intelligent navigation assistance for vehicle selection and service queries.
OUTCOME
One platform for the ownership journey
iGarage became a platform that handles the full arc of automotive ownership in one place: finding a vehicle, understanding its service history, booking maintenance, and transacting — buy or sell.
The Medusa integration gave the commerce layer a solid foundation without requiring a custom backend. The n8n AI integration added genuine utility — helping users who weren't sure where to start.
A consumer product built with the same care as enterprise software — because the stakes for users are just as high.
Organize around objects, not actions.
Centering the product on the vehicle — rather than on tasks like "book" or "browse" — made the information architecture far more intuitive.
Mobile is not a reduced version of desktop.
Designing mobile-first forced real hierarchy decisions. Desktop became an extension, not the original. The mobile experience is better for it.
Use proven infrastructure for commodity functions.
Medusa for commerce, n8n for AI workflows. Not reinventing what already exists lets focus go to the product decisions that actually matter.