Why I built Kontentpak: the content problem nobody talks about

I run a YouTube channel and a Telegram blog. Both in Russian, both about dev and AI.

Every time I covered a topic, the workflow was the same: write a YouTube script, then try to turn it into a Telegram post. Then write a description. Then headlines. By the third rewrite I'd lost all motivation.

The obvious move: use ChatGPT. But ChatGPT paraphrases. Feed it a script, ask for a Telegram post, and you get a shorter version of the same thing. Same structure, same points. That's not how good multi-platform content works. Your Telegram post should hook people from a different angle than your video. They should pull audiences between platforms, not repeat each other.

So I built Kontentpak. Two modes:

Content package: describe a topic, AI generates 10 ideas, pick one, select your platforms. Each module generates content with a different angle. The headlines module knows what the script says. The Telegram module knows what the headlines and script say. Each piece builds on previous ones.

Quick modules: need one thing fast (a headline, a script outline, a description). Skip the package, pick a module, go.

Current state:

• 15 AI modules

• Full pipeline wizard with image generation for covers

• History, favorites, export

• Payments live (YooKassa)

• Free tier: 11 generations/month

• Stack: Next.js, TypeScript, Supabase, OpenRouter

I spent more time tuning the AI prompts than writing the actual code. Each one is hand-crafted so the output doesn't read like generic ChatGPT. That was the hardest and most interesting part.

Revenue: $0. External users: 0. Built the whole thing solo.

Try it: kontentpak.space

If you create content for multiple platforms, I'd love to know: how do you handle it today?