Skip to content

MEG

Live · client work

A bilingual flagship site for a Doha furniture house.

A bilingual storefront for a Doha furniture house — designed and built by hand, shipped under a real domain.

Role
Design & build
Context
shenawy.group · Doha
Period
2026
Status
Live · client work

The brief

MEG is a furniture and interior-design business in Doha with a showroom, a large catalogue, and a brand that lives on craftsmanship. The site needed to work in Arabic first, not as an afterthought, and still give English-speaking clients the same experience.

The goal was not a generic catalogue. It was a digital extension of the showroom: browse by room, see the detail, understand the material, and start a conversation with the design team.

MEG interior with a carved fireplace and classic furniture
MEG showroom interior, Doha. Site imagery supplied by MEG.

What I built

I designed the visual system and built the site end to end: 27 pages across Arabic (RTL) and English, a 613-product catalogue with 638 images, projects, an atelier story, a material library, and a consultation flow.

  • Catalogue experience: category and subtype filters, bilingual search over English and Arabic names, URL state, load more, and quick view.
  • Lead pipeline: Flask API with Cloudflare Turnstile, a honeypot, SMTP alerts, SQLite/CSV storage, a protected admin dashboard, and CSV export.
  • Deployment: Cloudflare → nginx → static files plus a Flask service on systemd, with Let's Encrypt and daily backups.
  • SEO: localized metadata, hreflang, FurnitureStore/WebSite/FAQPage JSON-LD, a sitemap, and an llms.txt file.
MEG interior design project
A project page from the bilingual catalogue.

The details

The front end is hand-authored HTML5, CSS, and vanilla JavaScript with self-hosted Bodoni Moda, Manrope, Amiri, and IBM Plex Sans Arabic. There is no CMS and no checkout: the catalogue is inquiry-driven, which matches how the business actually sells.

The Arabic root is the primary experience, with the English mirror under /en/. That means RTL layout, Arabic-first typography, and localized metadata are product requirements, not a translation layer.

What I learned

RTL is a design constraint, not a mirror. Type, spacing, and carousels all needed to be designed twice, and the bilingual search needed to understand both naming systems.

A hand-authored stack was the right call for this site: it loads fast, has no build step, and can be handed over without a framework upgrade treadmill. The lead API is the only moving part, and it is small.

Next case study

Matflow

Experimental rows in, ranked candidates out.

Read the case study

Summer 2027

Hiring for an AI or software internship?

I am looking for a team where evaluation, product engineering, and research meet. Available from May 2027.