<aside> 💡
The problem: Creating platform-specific visuals meant generating or resizing an image by hand for each platform's aspect ratio, then posting it to the channel manually with no review step in between.
What it looked like before: Requesting an image and formatting it correctly were two separate manual tasks, getting the dimensions wrong meant wasted generations or awkward cropping, and nothing stopped an AI-generated image from going straight to a live channel unchecked.
What we built: A Telegram bot that detects the target platform from a single keyword, generates the image in the right aspect ratio automatically, and only publishes after the requester explicitly approves the preview.
Result: What used to be a two-step manual process (generate, then resize) is now one text command, with three distinct formats handled through a single interface. Every image passes through a human approval gate before it reaches the channel, so nothing gets published unchecked.
</aside>
An automated Telegram bot that generates AI images with platform-aware formatting and publishes to a channel only after human approval, built with n8n and OpenAI's image API.
Context:
Problem:
<aside> 🧠
The system receives an image request via Telegram, detects the target platform from a keyword in the message, generates the image in the correct aspect ratio, sends it back for review, and only publishes to the channel after the user explicitly approves it.
Tools used: n8n, Telegram Bot API, OpenAI (GPT-image model), Switch/Merge routing logic.
</aside>
Solution logic: