CASE STUDY · /deal-on-property

Deal On Property — Developer Build

A production real estate marketplace I built solo as the developer — Next.js 16 / React 19, server-first architecture, deep SEO surface, geo-aware routing, owner/agent workflows, companion Android app.

ROLE

Lead Developer · Architect · Solo Engineer

TIMELINE

2025 – Present

TEAM

1 (solo, AI-augmented)

STATUS

Production · Web + Android live

LIVE

↗ Website ▶ Play Store · Android

Overview

A real-estate marketplace I built solo as the developer — a platform going head-to-head with 99acres, Magicbricks, and Housing.com in the Indian market. Owners and agents list properties; buyers and renters discover, save, compare, and inquire. Web at dealonproperty.com, companion Android app live on Google Play.

Server-first architecture on the latest Next.js 16 / React 19. Every route is RSC by default — the property detail page, search results, city landing pages, and blog all render on the server with structured data baked in. Auth-gated dashboards (my-properties, leads, favorites, billing) handle the personalized surface.

The competitive bet is SEO: every city + locality + project becomes its own indexable surface, instantly notified to Google + Bing + Yandex via IndexNow on every listing approval — so a new listing appears in search the same day it's published.

The problem

Indian property portals optimize for inventory volume and ad revenue, not user experience. Listings are stale, owners are buried under agent calls, and search UX hasn't evolved in a decade. Building from zero on a modern stack means starting from clean fundamentals: fast pages, honest verification, structured booking instead of open-ended inquiry forms.

What I built

Next.js 16 App Router with React 19 Server Components — every public page is server-rendered, hydration only where interactivity is needed
5-step property wizard (basics → details → location → media → review) with HLS video upload, image gallery, and MapLibre location picker
Geo-aware routing: middleware detects city via Vercel/Cloudflare edge headers, sticky cookie preference, and optional precise-geolocation upgrade — never prompts on page load
URL structure that maps to SEO intent: /properties/[country]/[city] as route segments; locality, type, price as query params (avoids static-param explosion)
Multi-tier sitemap: pages.xml + locations.xml + blog.xml + per-property sitemaps under sitemap-index — regenerated on every listing approve/reject via revalidate webhook
IndexNow integration: single key, pinged to Bing/Yandex/Seznam on every listing change — same coverage as Google's Indexing API in one HTTP call
JSON-LD structured data: RealEstateListing, Product, BreadcrumbList, ItemList per page — Google rich results, Bing rich previews
TanStack Query for client data, Zustand for client state, Zod for schema validation, React Hook Form for the wizard — type-safe end-to-end
Admin panel: moderation queue, coupon engine, employee accounts, inquiry management with status workflows
Billing surface with checkout result toast, plan management, coupon redemption — wires into the same auth cookie as the user dashboard
Cron-driven sweepers: property expiration, saved-search alerts (planned), stale-listing nudges to owners
MDX blog system for city guides and SEO content surfaces — auto-aggregated into the sitemap

Tech stack

Next.js 16React 19TypeScriptTailwind v4TanStack QueryZustandZodReact Hook FormMapLibre GLHLS.jsFramer Motionnext-mdx-remoteVercelIndexNow APIGoogle Indexing API

Outcome

Live in production at dealonproperty.com. Every new listing is server-rendered with structured data and pinged to Google + Bing + Yandex within seconds of approval. Core Web Vitals green on property detail pages. The platform is the live proof of what a single AI-augmented engineer can ship at the scale of a 10-person team.

Want this in your stack? Let's talk.

Email Athar Back to all projects