FAVORIT: card generation (TG bot v1)

About the client

FAVORIT is JimmyNeuron’s e-commerce product line. The client needed a fast way to ship marketplace cards.

We started not with web, but with a Telegram bot: validate demand and the generation pipeline with real sellers. This case is stage 1 only (bot, v1). The full web app is stage 2: /cases/favorit-infocards

The challenge we started with

Test demand for card generation without a heavy web app
Give sellers an entry point through the familiar Telegram
Assemble the generation pipeline "as is" - image quality is still raw
Do not mix stage 1 (bot) with stage 2 (web) - different products in maturity

This is stage 1 · Telegram bot only

Version 1 of the project is a Telegram bot, not a full web service. The user generates cards right in the chat. The goal of the stage: speed and hypothesis testing, not a "pretty production".

Generation quality in the prototype

The gallery below shows real examples from v1. The design is weak: template infographics, crooked texts/logos, unrefined scenes. This is what the first iteration looked like - honestly, without retouching to look "already perfect".

Why a bot at all

Telegram removed the entry barrier: no registration on a separate site is needed. Sellers tried generation, and we watched what breaks in the pipeline and where quality hits its limit.

Project continuation → stage 2

The bot is not the end. Next we built a full web application with a different product maturity. Read the continuation and see the result we achieved: /cases/favorit-infocards

Gallery

FAVORIT: Telegram bot for marketplace cards - screenshot 1

FAVORIT: Telegram bot for marketplace cards - screenshot 2

FAVORIT: Telegram bot for marketplace cards - screenshot 3

FAVORIT: Telegram bot for marketplace cards - screenshot 4

How it works

User scenario

Opens the FAVORIT Telegram bot
Starts generating a product card
Gets an image in the chat (quality at the v1 prototype level)

Team / admin scenario

Sees what users actually generate
Records weak spots in design and the pipeline
Prepares the requirements for stage 2 - a full web app

Technology

Stage: 1 of 2 · version 1
Format: Telegram bot (not web)
Quality status: Prototype · weak design
Continuation: Web v2 → /cases/favorit-infocards

Timeline: TG MVP → web with quality expansion

Result and business effect

Launched a Telegram bot for card generation - FAVORIT stage 1 (v1). Continuation - full web service (stage 2): /cases/favorit-infocards.

A fit if you need

You need a quick MVP in Telegram before a heavy web app - AI agents
You want to test content generation on real users - AI agents
You plan to grow the bot into a full web product - Web applications

Frequently asked questions

Questions about this type of project, the same ones clients ask before a similar build.

Will the cards pass marketplace moderation?

We build each marketplace’s requirements (required fields, length limits, banned wording) into the pre-upload check. The final call is always the marketplace’s, though, and the rules change. When they do, we update the templates.

Can the agent make up an attribute?

We set it up to write only from your data. If the sheet has no material or size, the field stays empty and the card is flagged for a person. No card goes live without the content manager’s approval.

Do you guarantee more sales or better rankings?

No. We speed up card production and keep it consistent. Rankings depend on price, reviews, ads and the marketplace’s algorithms.

Can we start small?

Yes. We started a similar project as a Telegram bot, tested generation with real sellers, found the weak spots and only then built a web service.