The creator sends a finished video to a Telegram bot, and the system transcribes it, writes the title, description, and hashtags, and publishes it to several platforms on a schedule. No manual uploads, no formatting posts by hand
Global Import is a wholesale company based in Krasnoyarsk, Russia, that works with a large roster of UGC creators. These bloggers shoot video reviews daily and hand them off to be published on several platforms at once — 10+ pieces of content every day.
Every platform has its own rules and formats. For a video to get seen, it needs a title, a description, and hashtags — and it has to go out at the right time. With this many creators and videos, publishing turns into a job of its own.
Before, creators uploaded each video to every platform by hand, one by one, writing a fresh description every time. That repetitive work ate into time for shooting and editing.
The goal: creators do one thing — send the finished video to the bot — and everything else happens automatically.
Creators just send the finished video to the Telegram bot, and it’s posted to every platform automatically.
Titles, descriptions, and hashtags are generated from the video’s transcript in the creator’s own style, which is configured once.
Videos go out in the 5, 7, and 8 p.m. Moscow time slots and don’t overlap: each time is recorded in the schedule spreadsheet.
Once a video is live, a report lands in Telegram, so the creator can see right away that the post went through. Access to VK (Russia’s largest social network) refreshes automatically, so there’s no need to re-authorize.
Here’s the path of one video, step by step: from a message to the bot to going live on every platform.
The creator sends a video to the Telegram bot. That instantly triggers the workflow: the system picks up the file and starts processing. There’s no start button to press.
The audio track is pulled from the video, and a speech recognition model transcribes it. The transcript is the basis for the copy: the AI model works from what the creator actually says, not generic phrasing.
From the transcript, the AI model writes the title, description, and hashtags. Writing rules live in the settings and are tailored to each creator: tone, length, and must-have elements. If the video mentions a product, a dedicated rule kicks in and adds a marketplace link to the description.
The system finds the next open slot in the schedule spreadsheet — today, or tomorrow if today’s slots are full — and books it. At the scheduled time, the video goes out to TikTok and YouTube through a publishing service and to Telegram through the bot, and it’s uploaded to VK through the API as a community post.
Also used: FFmpeg — pulls the audio from the video
Creators now have just one job: send the video to the bot. Everything else runs automatically.
A similar setup works for anyone who regularly produces short videos and posts them across several platforms.
We build solutions like this around your needs: turnkey AI products, AI agents and automation, integrations, and data.
Our servicesTell us about your project — we’ll talk it through and show you how to make this work for your business
We use cookies and Yandex Metrica to improve our website. Privacy policy
