Your own delivery app or an aggregator in Uzbekistan?

The short answer

Use an aggregator when you need to be found. Build your own channel when you already have customers who order again, enough volume to cover fixed monthly costs, and a reason to own the relationship: your prices, your promotions, your data.

Most businesses end up doing both. The aggregator brings new customers; a direct channel—a Telegram bot, a web ordering page or an app—brings regulars back without a commission on every order.

Who delivers in Uzbekistan today?

The market is concentrated in a few platforms. Yandex Eats launched in Tashkent in March 2023, and in October 2024 Yandex bought part of Express24’s assets while the Express24 app kept running. Uzum Tezkor grew quickly; in April 2025 the competition regulator recognised it as dominant, when it was present in 19 cities.

The field has also narrowed. Wolt stopped serving Uzbekistan after 4 March 2026, and Bolt’s food delivery was only discussed as a possible pilot in 2025. For a restaurant or shop, that means a small number of partners setting the terms.

What does an aggregator really cost?

The platforms do not publish their commissions for Uzbekistan; the rate is set in each partner agreement. A 2022 column by the head of a local courier service put the average at about 20% of the order value. Treat that as an old estimate and ask for the current figure in writing.

The commission is only part of the price. On an aggregator the customer, the reviews and the order history belong to the platform; your brand sits next to competitors; and promotions follow the platform’s rules. In return you get demand you could not buy as cheaply yourself. The decision is arithmetic: monthly orders times commission, against what your own channel costs to build and run.

What is a delivery app made of?

People picture one app. A working delivery service needs three. The customer app shows the menu or catalogue, takes the address and payment, and shows where the order is. The courier app lists assigned orders, guides the route, records each status change and the cash collected. The admin or dispatcher panel lets staff accept orders, assign couriers, change prices and zones, and see what went wrong.

Behind them sits a backend that keeps the three in sync, plus the integrations: maps, payments, SMS or Telegram sign-in, notifications, and often the point-of-sale or accounting system. Most budget surprises come from the courier app and the dispatcher panel, because they are invisible in the first sketch.

Maps and addresses: the hard part

Many customers describe a location by a landmark, not a street number, and a pin on the map is often more reliable than a typed address. Plan for a map pin, a landmark field and a phone call as the fallback.

Choose the map provider with your cities and your use in mind. Google Maps Platform gives a free monthly allowance for each service and then charges from about $5–7 per 1,000 requests for maps, geocoding and routes. Mapbox includes 50,000 web map loads, 25,000 mobile users, and 100,000 geocoding and directions requests a month free. Yandex’s free terms forbid using the map to track vehicles or couriers, so live courier tracking needs a commercial licence; its commercial geocoder starts at 195,000 rubles a year. 2GIS has detailed maps for cities such as Tashkent, Samarkand and Bukhara, so check coverage for every city you serve.

Payments, cash and sign-in

Payme lists 2.0% for e-commerce payments with billing and 1.5% without billing. Click lists 1.5%–2.5% for Uzcard and Humo payments on websites, and offers a Telegram food-ordering bot at the same range. Plan time for merchant onboarding with either provider.

Cash on delivery is still common, so the courier app should record cash per order and the admin panel should reconcile it at the end of each shift. For sign-in, SMS codes are simple: Eskiz lists 95 soum per message for companies. Telegram sign-in avoids the SMS cost when your customers already order through Telegram.

Start with Android

StatCounter put Android at about 82% of mobile use in Uzbekistan in September 2026. Couriers almost always work from Android phones, so the courier app can start on Android alone. For customers, an Android app, a Telegram Mini App or a fast web ordering page can come first, with iOS added when the numbers justify a second store and its $99 yearly developer account.

Could Telegram be the first version?

Often, yes. A Telegram bot or Mini App can show the menu, take the address and the payment, and send status messages, while a simple courier app and an admin panel handle the operation. Customers do not install anything, and you learn which areas, hours and items matter before building a full app.

The limits are the ones in our Telegram guide: the bot cannot message people who never started it, and complex browsing is easier in a Mini App or a native app.

What does it cost to build?

No studio in Uzbekistan publishes a price for a delivery app, so the closest published figures are a guide. Innosoft lists mobile apps at 20 million soum for one platform with five to seven screens, 30 million for iOS and Android with ten to fifteen screens, and 50 million and up for larger products; its online stores run from 15 to 40 million. OneDev lists custom MVPs from $5,000. At the Central Bank rate of 12,056 soum per dollar on 17 October 2026, 20–50 million soum is roughly $1,660–$4,150.

A delivery service is three apps and a panel, so it usually costs more than a single-app package. White-label delivery platforms are the alternative: Delever, for example, sells a ready platform by monthly subscription from about 1–1.3 million soum. You trade ownership and flexibility for a faster start. AppX quotes in line with the local market, and the scope decides the price.

What does it cost to run each month?

The build is a one-time cost; the service runs every month. Maps and routing grow with orders and couriers, payment fees grow with revenue, SMS grows with sign-ins, and hosting and support grow with how much the business depends on the system. The table below shows what drives each cost and the published prices to start from.

A sensible first version

Start with one city or one zone, one order flow and a short list of items. Customers order through a Telegram bot, a web page or a simple app. Couriers use an Android app that shows the order, the route and the cash to collect. Staff use one panel to accept, assign and close orders. Manual steps are fine at first: a dispatcher can assign couriers by hand.

Measure orders per day, delivery time, failed deliveries and repeat customers. Automate what repeats—courier assignment, delivery zones and fees, promotions—only after the numbers show where time is lost.

Your own app and aggregators together

You do not have to choose one for ever. Keep the aggregator for discovery and give regulars a reason to order directly: faster confirmation, a loyalty discount, or free delivery on direct orders. Read the partner agreement first; some contracts limit how your direct prices may differ from the platform’s.

Over time, the share of direct orders shows whether the own channel is paying for itself.

Questions to answer before you ask for a quote

Which cities and zones will you serve first, and in which hours? How many orders a day do you expect in the first three months? Who delivers—your own couriers, a partner, or both? Which payments do you accept, and how will cash be reconciled? Which map provider fits your cities and courier tracking? What must connect to your existing systems? Who will run the dispatcher panel every day?

Written answers turn a vague estimate into a quote you can compare between studios.

Where does AppX fit?

AppX Tech Group is a product studio in Urgench, Uzbekistan. We design and build customer apps, courier apps and admin panels as one system, in Uzbek, Russian and English, and we quote in line with the local market. Our logistics page describes how we approach the operational side.