CALL USEMAILWHATSAPPGET QUOTE
Marketplace

ACE Krt: Architecting a Multi-Vendor Marketplace for 25K+ Products

Priyanka Ghosh

Priyanka Ghosh

Chief Technology Officer · India

Chief Technology Officer at Aprodence, setting the engineering direction and keeping every platform scalable, secure, and fast. She has led teams through re-architectures, cloud migrations, and high-traffic launches, and believes the best technology is the kind users never have to think about.

Mar 19, 202610 min read
ACE Krt: Architecting a Multi-Vendor Marketplace for 25K+ Products

A marketplace is many products in one: storefront, vendor portal, payments, and admin. Here's how we architected ACE Krt to scale past 25,000 listings.

25K+
Products Listed
800+
Active Vendors
< 200ms
Search Latency

Multi-Vendor from Day One

We designed the data model around vendors, not just products, so onboarding new sellers was instant and self-serve.

A flexible catalog schema in MongoDB let vendors define custom attributes without breaking search or filtering.

Search & Filtering at Scale

With tens of thousands of products, discovery is everything. We built fast, faceted filtering that stays snappy even as the catalog grows.

Payments & Admin

Integrated payment gateways handled split payouts to vendors, while a powerful admin panel gave the team full control over orders, disputes, and reporting.

ReactNode.jsMongoDBTailwind CSS
PG

Written by

Priyanka Ghosh

Chief Technology Officer, Aprodence

FAQ

Questions this build raised.

The decisions, trade-offs, and gotchas behind the work — answered straight.

Still curious? Talk to us

Database LIKE queries fall over well before 25K. You need a dedicated search index with facets precomputed, so filtering by vendor, category and attribute is one lookup instead of nested joins. The real cost is reindexing: every vendor price edit must propagate in near real time without rebuilding the whole index.

Have a project like this?

Let's turn your idea into the next case study. Tell us what you're building.