Telegram bot, Mini App or app? A guide for Uzbekistan.

The short answer

Start in Telegram unless the product needs something Telegram cannot give. Build a bot when the job is a conversation: a question, a form, a status, a reminder. Build a Mini App when people need screens: a catalogue, a calendar, a cart, a dashboard. Build a native mobile app when the product depends on offline work, background activity, deep device features, or a place on the home screen that people open every day.

The three are not rivals. A bot is usually the door, a Mini App is the room behind it, and a native app is a second building you add when the first one is full. The useful question is which to build first, and how to avoid rebuilding when you add the next.

Why is Telegram the default starting point in Uzbekistan?

Because your customers are already there. Estimates of Telegram’s audience in Uzbekistan range from 18 to 25 million users, in a country that DataReportal counts at 33.1 million internet users. Telegram itself reports more than one billion monthly active users worldwide. A product inside Telegram asks for no download, no new account and no storage space.

It also sidesteps the platform split. StatCounter put Android at about 82% and iOS at about 18% of mobile use in Uzbekistan in September 2026. A native product needs both to reach everyone. A bot or Mini App runs the same way on both from the first day, with no store review between you and a release.

What can a Telegram bot do?

A bot is a Telegram account run by your software. Through the Bot API it receives messages and button presses and answers with text, photos, files, menus and inline buttons. It can take payments, work in groups, and send a message the moment something changes on your side.

That makes a bot strong at short, structured exchanges: booking a slot, tracking an order, collecting feedback, answering common questions, alerting staff. Feekr, in our portfolio, uses this pattern. A customer scans a QR code at the table or counter and rates the service in Telegram, with nothing to install.

Where does a bot run out?

A chat is a line, and some tasks are not linear. Browsing forty products, comparing options, editing a long form or reading a dashboard is awkward as a sequence of messages. People lose their place, and every correction means scrolling back.

Bots also follow rules that shape the product. Telegram does not let a bot start a conversation: the person must message it or open it first. Mass sending is limited to about 30 messages a second unless you pay for broadcasts, and to roughly one message a second inside a single chat. A bot is a service channel, not an advertising channel.

What does a Mini App add?

A Mini App is a web application that opens inside Telegram—from the bot’s menu button, an inline button or a direct link. It is built with ordinary web technology, so it can show real screens: catalogues, calendars, maps, carts, charts and multi-step forms.

Telegram has steadily widened what a Mini App can reach. The official documentation covers full-screen mode, a shortcut on the phone’s home screen, geolocation, biometric confirmation, a QR scanner, haptic feedback, motion sensors, and storage on the device and in the cloud.

The app also knows who opened it. Telegram passes signed user data that your server verifies, so there is no password to create. PrepGo, in our portfolio, uses the same idea on the web: learners sign in with Telegram.

Where does a Mini App run out?

It lives inside another company’s product. If Telegram changes a rule, an interface or a payment policy, you adapt on its schedule. People who do not use Telegram cannot use your product unless you also publish it as a website.

A Mini App is active only while the person is looking at it. It cannot work in the background, and its notifications arrive as messages from your bot instead of system push notifications. Long offline sessions, heavy graphics, continuous location tracking and Bluetooth devices remain native territory. Storage is deliberately small: Telegram’s documentation sets a limit of 5 MB of device storage per user for each bot.

What does a native app add?

A native app is installed from the App Store or Google Play and runs on the phone itself. It can work without a connection, run tasks in the background, send system notifications, and use the camera, sensors, Bluetooth and files more deeply than a web view can. It also earns a permanent icon on the home screen and a listing people can find in the store.

The price is process. You need an Apple developer account at $99 a year and a one-time $25 Google Play registration. Every release passes store review; Apple says it reviews at least half of submissions within 24 hours and 90% within 48. You maintain two platforms, follow two sets of guidelines, and ship updates when the operating systems change.

How do payments work in each?

Telegram separates two kinds of sales. For digital goods and services—subscriptions, content, in-app features—bots and Mini Apps must take payment in Telegram Stars, Telegram’s own currency. For physical goods and real-world services, such as food, clothing, tickets or a haircut, Telegram’s Payments API connects to outside payment providers, and Telegram charges no commission of its own.

In Uzbekistan both major providers document this route. Payme publishes instructions for connecting a bot and lists Telegram bot payments in its e-commerce tariff at 2.0%. Click lists payments through a Telegram bot at 1.5%–2.5% and states that it charges no connection or subscription fee. Plan time for the provider’s merchant onboarding before the first real payment.

Native apps follow store rules instead. Digital goods sold inside an iOS app go through Apple’s billing at a 30% commission, or 15% under its Small Business Program. Google Play’s standard service fee is 15% on the first $1 million a year and 30% above it in most markets. Physical goods and services can be paid through a local provider such as Payme or Click.

What does each option cost in Uzbekistan?

Published prices give a useful floor. Innosoft lists Telegram bots from 3 to 20 million soum: 3–5 million for a simple information and order bot, 5–8 million with Click or Payme payments, 8–15 million for an AI chatbot, and 10–20 million for a bot tied to a CRM or ERP. SoftWhere lists information bots at 3–5 million and Mini Apps at 10.5–35 million soum. UZNEO’s guide puts a simple bot at 4–12 million and one with payments, a catalogue and an admin panel at 20 million or more. OneDev lists business chatbots from $500.

At the Central Bank rate of 11,847 soum per dollar on 9 October 2026, that is roughly $250–$1,700 for a bot and $900–$3,000 for a Mini App. Published starting prices for a mobile app run from a few million to about 40 million soum, as our cost article shows, and real multi-role apps cost well beyond the starting price.

AppX works at local prices too. For businesses in Uzbekistan we quote in line with the local market, and the scope decides the number: a simple bot is priced as a simple bot, and a product with several roles, payments, an admin side and integrations is priced as a product. Send us the scope and we will give you a figure you can compare.

What actually takes the time?

Rarely the part the customer sees. A bot’s messages or a Mini App’s screens are the visible tenth. The rest is the work behind them: the admin panel where staff see orders and change prices, the connection to a payment provider, the link to an accounting or delivery system, and the rules for refunds, cancellations and mistakes.

Language multiplies the effort. A product in Uzbek and Russian needs every message, button, error and notification written and tested twice, and often a third time in English. Published timelines reflect this. Innosoft quotes 7–10 days for a simple bot and up to 30 for a CRM bot, and OneSystem quotes three to six weeks for a Mini App with payment and CRM integration.

Three patterns from our portfolio

Feedback at the point of service. Feekr places a QR code where the customer already is. Scanning it opens Telegram, and the customer answers in about a minute. Asking for an app download at that moment would be asking too much.

Sign-in without a password. PrepGo lets learners sign in with Telegram. There is no password to forget, and the account is tied to an identity the learner already uses every day.

Sales conversations where customers already write. Seylo works in Instagram and Telegram, the channels where shoppers already ask about prices and delivery, and keeps a person in control of what the business promises. In each case Telegram is not the product. It is the shortest path between the product and the person.

How do you choose? Seven questions

Answer these before asking for a quote. One: can the main task be finished in five or six messages? If yes, start with a bot. Two: do people need to browse, compare or edit? That points to a Mini App. Three: must it work without a connection or in the background? That is native. Four: are you selling digital goods? Then check the Stars rule and the store commissions first. Five: are your customers on Telegram at all? Tourists and some corporate buyers may not be. Six: who will operate it each day, and what do they need to see? Seven: what would make you outgrow the first version?

If the answers are mixed, the usual order is a bot first, a Mini App when the bot starts to feel cramped, and a native app when usage justifies the upkeep.

When should you move from Telegram to a native app?

Move when the evidence says so. People use the product daily and ask for an icon on their home screen. The work happens where the connection is poor: drivers, field staff, warehouses. You need background tracking, Bluetooth devices or richer media. A partner or regulator expects a store listing. Or you want a channel you fully control.

Move for a reason you can name, and keep the Telegram entry. For many customers it will remain the fastest way in.

How do you build so you do not rebuild?

Treat Telegram as one client of your system, not as the system. Keep the business rules, data and payment logic in your own backend behind a clean API. The bot, the Mini App, a website and later a native app then become different doors into the same building.

Three habits protect you. Register the bot yourself, so the access token and the audience belong to your company and not to a contractor. Keep your own customer records, linked to the Telegram identity but not dependent on it, with a phone number or email where the customer agrees. And build the Mini App as a normal responsive web application, so the same code can serve people outside Telegram as a website.

Where does AppX fit?

AppX Tech Group is a product studio in Urgench, Uzbekistan. We design and build bots, Mini Apps, web and mobile products as one system, in Uzbek, Russian and English. If you are deciding where to start, send us the main journey in a few sentences and we will tell you which option we would build first, and why.