Dubai · Web Development
E-commerce website development in Dubai — built for how the UAE actually pays.
Your storefront wired to the payment methods, delivery rules and 5% VAT that UAE retail runs on, and joined to wherever your stock already lives. You finish with a live store, an operations handbook, and every gateway account in your own name.
- UAE gateways, not global defaults
- Cash on delivery, reconciled
- Gateway accounts in your name
What's included
Six things that have to work behind the buy button.
The storefront is the part everyone reviews before launch. These are the parts that generate the support emails afterwards.
A catalogue modelled before it is filled
Products, variants, options, bundles and the price rules sitting on top of them, decided as a data model rather than discovered one product at a time during content entry. The model is what a large catalogue eventually has to be rebuilt around, so it is settled first.
Checkout and UAE payment methods
Cards through Network International, Telr or PayTabs, Apple Pay, and Tabby or Tamara where instalments matter to your buyer. Each method is exercised end to end — successful payment, declined card, abandoned 3-D Secure, partial refund — because the failure paths are the ones customers actually meet.
Delivery, collection and cash on delivery
Shipping rules by emirate, weight or order value; collection where you have a shop; and cash on delivery treated as a real order state rather than a checkbox. An unpaid order has to reconcile against what the courier remits, and that reconciliation is built rather than left as a spreadsheet you inherit.
VAT, invoices and refunds
5% VAT configured in the tax rules rather than calculated in a template, and a tax invoice carrying your TRN instead of a styled order confirmation. Refunds and partial refunds reverse the tax as well as the money, which is the part that quietly goes wrong on most stores.
Stock and orders joined to what you already run
The store is rarely the only place stock exists. One system is named as the source of truth for stock, price and the customer record, data flows in one direction rather than two, and the behaviour when a sync fails is decided during architecture instead of during an oversell.
Card details you never hold
Card numbers are entered into the gateway's own hosted fields and never reach your server, which keeps the compliance surface small. Accounts, resets and order history are built with the same restraint: store what the business genuinely needs and nothing that becomes a liability later.
Who it's for
Three businesses this is the right build for.
If a hosted platform covers your requirements without argument, that is the simpler purchase — and we would rather say so now than in month three.
A retailer whose stock lives somewhere else
You already run a POS, an ERP or a warehouse system, and the website would become the fourth place a stock number is typed. The storefront is the straightforward half of this project. Deciding which system is allowed to be wrong is the hard half, and it is where the work starts.
Someone updates stock in two systems every morning.
A brand that has outgrown a hosted platform
The catalogue is fine; the rules are not. The checkout has to behave in a way the platform will not permit, the shipping logic does not fit the app that sells it, or the platform fee has become a line item you notice. Moving is a data project first and a design project second.
Your last three requirements each needed a paid app.
A B2B seller whose prices are not public
Account logins, contract price lists, minimum order quantities, quotes, and re-ordering from last month's order. None of that is unusual in UAE trade, and almost none of it is the shape a consumer storefront was built in. It can be bent into that shape; it is usually cheaper not to.
Two logged-in customers should not see the same price.
How we deliver it
Five stages, and money moves in the fourth.
Each stage ends in something you can click, place an order in, or refuse to sign.
01
The product and pricing model
Every attribute, variant, bundle and price rule written down with its tax treatment against it, then signed off. Changing the model later means changing the data as well as the code, which is why it is not a build-time decision.
02
Where the truth lives
Which system owns stock, which owns price, which owns the customer record, and which way data moves between them. Settled on paper while it is still an argument rather than a migration.
03
Storefront and checkout
Listing, filtering, product pages, cart and checkout, on a staging URL with the gateways in test mode. Weekly increments you can click through and place orders in, from the first week there is anything to click.
04
Money, tax and delivery in test
Every payment method run end to end, including the paths nobody demos: declined cards, an abandoned 3-D Secure step, partial refunds, a cash-on-delivery order refused at the door, and VAT reversed correctly on each. Shipping rules are checked against real addresses across the emirates.
05
Migration, launch and the handbook
Catalogue, customers and order history moved across, every old product URL mapped to its replacement, then a first live transaction placed and refunded on a real card before you take the keys. The handbook and the training happen here, not afterwards.
Stack
Boring technology, chosen so the next developer recognises it.
Nothing here is exotic. That is the point: an unusual stack is a lock-in mechanism dressed up as innovation.
- Front-end
- Next.jsReactTypeScriptTailwind CSSAstro
- CMS & commerce
- WordPressShopifySanityPayloadWooCommerce
- Back-end & data
- Node.jsPostgreSQLLaravelRedisPrisma
- Infrastructure
- VercelAWSCloudflareGitHub ActionsSentry
Deliverables
What you receive when the store goes live.
All of it in your accounts, and all of it written for the person who will run the catalogue rather than the person who built it.
- A live store, with the first real transaction placed and refunded before you take it over
- The operations handbook — adding a product, running a refund, reissuing a tax invoice, and what to do when a payment sits pending
- The product model as a document: every attribute, variant, bundle and price rule, with its tax treatment
- A test-order checklist you can re-run after any change — card, Apple Pay, instalments, cash on delivery, refund
- Gateway, courier and platform accounts opened in your name, with the credentials in your password manager
- A redirect map from every old product and category URL to its replacement
- The repository, and a README that runs the store from a clean clone against test gateway keys
- Recorded training for whoever runs the catalogue day to day
The handbook is the item that usually gets skipped, and it is the one that matters most in month two. The questions that follow a store launch are operational rather than technical — a refund that will not go through, an order stuck between paid and shipped, a payment that shows in the gateway but not in the store. Writing the answers down before launch is cheaper than answering them one at a time.
Pricing
What an online store costs.
AED 48,000
from
A store is not a Launch-tier purchase. A first store belongs in the Growth tier from AED 48,000, which carries a custom content structure, Arabic and a payment integration. Larger catalogues, B2B pricing and stock synced to an ERP sit in the Platform tier from AED 95,000. Prices exclude 5% VAT, and payment runs in milestones.
What changes the priceWhat that covers
- A written product and pricing model before anything is built
- Checkout wired to UAE cards, Apple Pay and instalments
- 5% VAT configured, with tax invoices issued in your name
- Delivery rules, and cash on delivery if you take it
- Catalogue and customer migration, with old product URLs redirected
- The operations handbook and recorded editor training
Retail and consumer brands
The catalogue moves constantly — new lines, sizes, colours, variants that sell out mid-campaign — and most of the traffic arrives on a phone, often on mobile data rather than office wifi. What the sector needs is a product page that stays fast under photography weight, filtering that does not reload the world, and instalments at checkout, because a UAE checkout offering only a card is offering less than the one beside it.
Wholesale and distribution
Prices are negotiated rather than published, orders repeat, and the buyer knows what they want before they arrive. That means account logins, contract price lists, minimum order quantities, order history you can re-order from in two taps, and stock that reflects the warehouse rather than last night's export. The person browsing is also rarely the person who pays, so approval and invoicing belong to somebody else in the same company.
An online store is one of the builds described under our full web development service and it is the one where most of the work sits behind the buy button rather than in front of it. If you would rather not own the infrastructure at all — hosting, patching, the compliance surface — the Shopify page is the closer fit, and the one to read first.
Questions
Asked before a store gets built.
A store starts at AED 48,000 — the Growth tier, which carries a custom content structure, Arabic and a payment integration. Larger catalogues, B2B pricing or stock synced to an ERP move it to the Platform tier from AED 95,000. The AED 16,000 Launch tier is priced for a marketing site, and a store is not one: VAT on every order and refund, delivery rules, and stock kept honest against whatever system already holds it are work a page count never sees. Prices exclude 5% VAT.
Network International, Telr, PayTabs, Tabby, Tamara and Apple Pay, plus whichever your bank has already approved you for. The merchant account is yours, and the commercial terms — rates, settlement period, chargeback handling — sit between you and the provider rather than with us. What we need to build is sandbox credentials. Approval timelines belong to the provider, so they are worth starting in week one.
Yes, and it is worth building properly rather than bolting on. Cash on delivery creates an order that is confirmed but unpaid, so the store needs a state for it, a rule for whether it holds stock, and a reconciliation against what the courier actually remits. Refusals at the door and returns to origin have to be reversible in one place, VAT included, or your accounts and your store drift apart within a month.
In the tax rules, not in a template. Consumer prices in the UAE are normally displayed inclusive of VAT while business buyers expect to see it broken out, and both need a tax invoice carrying your TRN rather than a styled order confirmation. Refunds and partial refunds have to reverse the tax as well as the money. Whether and when you register for VAT is a question for your accountant, not your developer.
Usually, if the system exposes an API — but the useful question is what happens when the sync fails, not whether it exists. One system is named the source of truth, stock flows in a single direction, and we agree in advance whether the store hides the product, accepts a backorder or blocks checkout when the two disagree. An oversell is a phone call to a customer either way; the design decides who makes it.
Products, variants, images, customers and order history migrate, and every old product and category URL gets a redirect to its replacement. What does not migrate is anything the old platform held on your behalf — saved cards, which live with the gateway, and app-specific data that never had an export. Those are listed before the work starts, because discovering one on launch weekend is how a migration becomes a story.
No — one store, two content sets, with the layout mirrored rather than the site duplicated. The cost is not the right-to-left work; it is the catalogue. Every product name, description, attribute, filter value and delivery message needs an Arabic version, and that is content somebody in your business has to own. Machine translation on a product catalogue reads exactly as badly as it sounds. Arabic is included from the Growth tier.
No. Photography, copywriting and catalogue entry are not services we sell, and pretending otherwise is how a launch date slips quietly. What we do is specify the shape the content has to arrive in: image dimensions and crop, the attributes every product needs, and a spreadsheet template for bulk import. Content is the usual cause of slippage, so that template goes out in week one rather than week five.
Ready to build something that lasts?
Tell us what you need. You will get a scoped response with an indicative price — from an engineer, not a salesperson.