Automotive research platform · Web + Androidlive
CarBikeSpecs
India's car & bike specs, prices and comparisons — 3,600+ indexed pages on an SEO-first Next.js site, an admin platform with automated data pipelines, utility tools and a mobile app.
- role
- CTO, brandMongo — built from scratch
- period
- 2025 – present
- 3,600+
- indexed pages
- 2,500+
- car pages
- 1,000+
- bike pages
- 7
- utility tools
- 3
- surfaces: site · admin · app
# the idea
Spec sites live or die on data quality and search traffic. The product is really a data pipeline — scrape, clean, verify, publish — with a fast, indexable website and an app on top.
# what I built
- Public site on Next.js App Router: 2,500+ car pages, 1,000+ bike pages, comparisons, blog and web stories.
- Utility tools — loan EMI, car & bike insurance, e-challan lookup, RC details, RTO details and dealer locator.
- SEO engineering: canonical URLs, structured data, sitemap index + news sitemap, IndexNow and llms.txt for AI crawlers.
- Edge caching and routing through a Cloudflare Worker in front of Cloud Run.
- Admin platform with scrapers for OEM and marketplace sources, review queues, CMS for blogs, stories and banners, and lead management.
- Data-quality toolkit: de-duplication, price reconciliation, unit normalisation, launch-date and identity-drift checks — every bulk fix backed up first.
- Expo / React Native app as a pure client of the site's API, with deep links, push notifications and in-app review.
# key decisions
- One API powers web and app — no separate mobile backend to keep in sync.
- Scrapers live in the admin, never the public site, so a scraping failure can't take traffic down.
- Every destructive data migration writes a backup file before it runs.
# stack
- Next.js
- TypeScript
- MongoDB Atlas
- Expo / React Native
- Cloud Run
- Cloudflare Workers
- Firebase