Website + mobile app redesign
Full redesign of the website to the new Figma designs, including Arabic right-to-left layout, replacing the earlier Chinese/English demo with ¥ prices and hard-coded cars.
- AI Product Engineering
- Forward-Deployed Services
Mingcheji
Mingcheji gives buyers outside China one platform to check a Chinese car's history, find parts that fit it, and browse listings — in six languages, with prices in USD. Pilot on a preview site; not yet in production.
Pilot — preview siteView on Google Play →
Mobile app interface
Snapshot
Production metrics & deployment details.
Audited environment metrics verified against live deployment.
- Client / Context
- Mingcheji — global car-history and parts platform for buyers outside China (website + mobile app)
- Engagement
- Full website redesign + new backend API: sign-in, accounts, free part checks, Stripe pay-first orders, coupons, order queue, GlobalVIN vehicle data
- Status
- Pilot — running on a preview site, not yet in production
- Technology
- Next.js 16, React 19, Tailwind CSS 4, next-intl · Node.js, TypeScript, Express 5, PostgreSQL, Kysely, pg-boss, Zod · Stripe Checkout, Google sign-in · GlobalVIN, che168 via Apify · Vercel, Render preview, Sentry
The challenge
What high-stakes problem had to be solved?
The business context, constraints, and risk surface that shaped every decision.
Buyers outside China had no easy way to check a Chinese car's history, find parts that fit it, or browse listings — the data exists in Chinese, priced in yuan (CNY), and written for China's own formats.
The client needed one global platform in six languages (English, Arabic, Russian, Spanish, Portuguese, Chinese) with prices in USD. An earlier front-end demo (Chinese/English, ¥ prices, hard-coded cars) has been fully replaced; Car Search and the history report pages run on sample data until the live data is connected.
Contributions
What GenAILabs worked on.
Factual breakdown of our direct engineering contributions.
Backend API
New backend API: sign-in (Google, optional password), accounts, free part checks, Stripe pay-first orders, coupons, and an order queue.
Vehicle data access
Access to the vehicle data provider (GlobalVIN, via the China Vehicle History backend) for car-history reports.
Car Parts
VIN or part-number lookup, a car's full parts catalogue, and OEM prices in USD.
System / solution
Technical design of the custom system.
Next.js website, Node.js/Express backend, GlobalVIN vehicle data, Stripe payments, and self-repairing background jobs.
System architecture
Next.js 16 website (Vercel) → Node.js/Express 5 API (Render preview) → GlobalVIN vehicle & spare-parts APIs via the China Vehicle History backend; che168 listings via Apify for Car Search; Stripe Checkout payments; Google sign-in; PostgreSQL + Kysely + pg-boss job queue; Sentry monitoring. Reports show only provider-returned data, with "no data" states for empty sections.
Flywheel impulse
Deployment, integration, and real-world use.
Verified real-world deployment, maintenance, and scale protocols.
Pilot running on a preview site (Vercel website, Render API preview). Changes ship without a release: prices, limits, coupon rules and call costs are settings stored with history and picked up within 30 seconds. Production target is a VPS with Docker (proxy, API, worker, Postgres, backups) — not built yet.
- Integration pattern
- GlobalVIN vehicle and spare-parts APIs; che168 listings via Apify (chosen for Car Search); Stripe Checkout payments; Google sign-in
- Scale protocol
- At most about 5 GlobalVIN calls at once (paid orders first); circuit breaker pauses calls when GlobalVIN fails; daily budget stops spending past a cap
- Maintenance mode
- Prices, limits, coupon rules and call costs are settings stored with history, picked up within 30 seconds; background jobs retry deliveries for up to 24 hours
Outcomes / evidence
Audited metrics proving system efficacy.
Every metric listed is verified and sourced from the Mingcheji case data.
Verified evidence
Six languages on every page
next-intl — English, Arabic, Russian, Spanish, Portuguese, Chinese — with full Arabic right-to-left layout
Verified evidence
No lookup paid for twice
Results are stored; a search that finds nothing issues a 60-day coupon
Key outcome
Car Parts works end to end
Sign-in, free checks, Stripe payment, order queue, and a translated whole-car parts catalogue
Capabilities involved
How core competencies were applied.
AI Product Engineering and Forward-Deployed Services were both demonstrated in this pilot.
AI Product Engineering
Full website redesign to new Figma designs with Arabic RTL; new backend API and Car Parts lookup with OEM prices in USD — all demonstrated in the pilot.
Forward-Deployed Services
Pilot running on a preview site (Vercel website, Render API preview) with settings-based configuration picked up within 30 seconds and self-repairing background jobs.
Related work
None — Mingcheji is the first in this domain.
No related case studies exist yet for the Mingcheji project.
Let's talk
Have a similar challenge?
Let's discuss the problem you're trying to solve.
Talk to our team
