A WordPress-powered e-commerce store built with zero bloat. Custom plugins, Gemini AI product descriptions, and organic sales before launch day.






Every build is a list of choices. Here is what we chose, what we turned down, and why.
Budget constraints and client team needed self-service product management. WordPress gives the foundation; custom plugins give the performance.
Every plugin is attack surface and bloat. Custom code = full control, no wasted resources, auditable security.
Gemini API is nearly free for this use case. Batch processing keeps costs minimal. Quality matches GPT-4 for structured formatting tasks.
Fast pages + clean structure + quality content from day one = organic traffic from day one. No bolt-on SEO fixes needed.
EMEM Trading sells consumer products across multiple physical locations in Sri Lanka. They already move inventory through e-commerce marketplaces, and they move it well. But marketplace dependency means commission eat margins, you don't own the customer relationship, and you can't control the experience.
They knew they needed their own website. But here's the honest question every marketplace-dependent seller faces: is the D2C channel worth the investment before you've proven buyers will actually convert on it?
They didn't want to spend a fortune finding out.
We're a team that builds on Next.js, React, Laravel, Flutter. Recommending WordPress for a client project might seem like a step backward.
It wasn't. It was the right tool for this specific constraint set:
• Budget was lean: custom headless builds cost more
• Client's team needed to manage products themselves post-launch
• Time to launch mattered: WordPress + WooCommerce gives you a working commerce foundation out of the box
• The real value wasn't the platform: it was what we built on top of it
We agreed on a hard constraint with the client: only 3 core third-party plugins. Everything else gets built custom.
The 3 we kept:
1. WooCommerce: commerce engine (non-negotiable)
2. A lightweight SEO plugin: for sitemap and meta management
3. A caching plugin: for page-level caching
Every other feature (product image handling, location-based taxonomy, bulk operations, description formatting) was built as a custom plugin.
The most impactful custom plugin we built wasn't a visual feature. It was invisible to end users but transformational for the client's operations.
The problem: Products arrive with raw, inconsistent descriptions, sometimes just a manufacturer's spec sheet pasted in. Writing unique, SEO-optimized descriptions for hundreds of products is time-consuming and expensive.
The solution: A custom WordPress plugin that connects to Google's Gemini API. Paste a raw description, click generate, and get a clean, well-structured, SEO-aware product description.
The budget-smart part: We built it as a batch process. Select 50 products, hit 'Generate All', and the plugin processes them in a background queue. This keeps API costs nearly zero: we're talking pennies per hundred products. For individual products where speed matters, there's also an instant-generate option.
Here's the result that matters most:
Before EMEM Trading even publicly announced their website launch, they started receiving organic orders from Google Search. Not from ads. Not from social media. From people searching for products and finding emem.lk.
This happened because:
• The site was fast (PageSpeed 90+) so Google crawled and indexed it quickly
• The AI-generated descriptions provided unique, keyword-rich content on every product page
• The clean site structure with proper taxonomy meant Google understood the site hierarchy
• No bloat meant no render-blocking resources slowing down the page
“We didn't expect sales before we even launched. The site just started working on its own. That's when we knew the D2C channel was real.”

A German financial research and education brand rebuilt off Beehiiv entirely: website, financial dashboard, custom CMS, custom email platform, and a Telegram-bot-run premium community, with every subscriber and every article carried over.
A licensed Qatar ride-hailing app for Hello Limousine, a limousine company running thousands of drivers and cabs, built after other vendors left them with an app they didn't fully own. Architected from day one to launch in a new GCC country in minutes, not months.
A Doha-based Web3 social network built for Ismail, a top-3 Binance Live streamer with 150,000+ followers, so his community had a legitimate crypto project to join instead of another token to speculate on. Built in 90 days, run on retainer for two years, acquired by Cryptala.

You're not hiring a company. You're hiring a person who signs their name to it.
Rikaz Zarook, Founder & CTO. Thirteen years in software. Two builds a month, that's the whole capacity.