A Mini App is a web page that opens inside Telegram in a built-in browser. A plain bot is a conversation: messages, buttons beneath them, commands. Technically these are different things, and they cost three to four times apart.
What a button-based bot can do
A branching dialogue. The user taps a button, the bot replies with a message and shows the next set of buttons. This is how requests get collected, how appointment booking works, along with notifications and simple orders from a short menu.
Its strengths: it works instantly, needs no loading, looks the same on any device, and is familiar to everyone.
The ceiling arrives where you need to show a lot at once. A catalogue of two hundred items paged through buttons is painful to browse. Filters across four parameters turn a conversation into an interrogation. A basket that has to be edited forces the user to hold the state in their head.
What a Mini App can do
Everything an ordinary website can: a grid of cards, filters, a map, a calendar, a form with validation, animation, a basket you can edit.
Plus something a website lacks: the user is already signed in. Telegram passes the app signed data about the user, and registration does not exist as a step. For a shop this removes the main drop-off point.
Plus payment in two taps via a saved card.
Four signs you need a Mini App
More than thirty items in the catalogue. A button menu stops coping sooner than it seems.
The user chooses across several parameters at once. Size, colour, deadline, price. In a dialogue this is four sequential questions with no way to go back.
You need a map, a calendar, or a chart. A picture cannot replace them, and text is awkward.
The user comes back regularly. A personal account with history, balance, statuses. Scrolling order history in a chat is impossible.
If none of these apply, take the bot. It is cheaper, launches faster, and breaks less often.
What a Mini App cannot do
Send messages on its own. The app lives only while it is open. All notifications are still sent by the bot, which is why a Mini App almost always comes paired with one.
Work offline. It is a web page, with everything that follows from that.
Open from search. There is a single entry point: through the bot or a link.
On performance
Telegram's built-in browser on weak Android devices is slower than plain Chrome. A heavy interface that opens in a second on desktop will take three to assemble there.
The practical consequence: keep the first screen light and load the rest as the user scrolls. Animation libraries, three fonts, and full-resolution images are the fastest way to get an app that people close before it loads.
What the switch costs
A button-based bot is ₽40,000–70,000. A Mini App with a catalogue and payment is ₽250,000–450,000. The difference is in the volume of work: an app has design, markup, interface state, and a separate API layer that a bot simply does not have.
The sensible order for a new product is the bot first. It shows whether people use your scenario at all. After a month the data reveals exactly where the button interface gets in the way, and the Mini App gets built for a known problem.
What stays shared
The server side. The database, CRM integrations, payments, access rights: all of this lives on your server and does not depend on whether the user sees buttons or a screen. If the bot was designed with a proper API, adding a Mini App on top of it costs less than building from scratch, since only the interface has to be rewritten.