My name is Igor, I'm 34, and I've been hosting an entrepreneurship podcast for three years. YouTube: 180,000 subscribers. Telegram: 60,000. Instagram: 45,000. A year ago, I worked solo, putting out about two long episodes a month and four Telegram posts a week. That was my limit. I hit a point where it was either hire a four-person team for 200k in monthly expenses or quit content altogether.
I picked a third option: I kept myself, and built a content factory using Quantium AI. Now I make four long episodes a month, 20 shorts, 20 Telegram posts, and a bunch of Pinterest pins. All solo, no burnout, just four hours of work a day. Below, I'll walk you through my workflow step-by-step, explain which models I use for what tasks, and tell you how many credits an active blogger burns through in a month.
Step 1. Post Ideas: 50 in an Hour with ChatGPT
I used to get stuck on "what to write about today." I'd open notes, scroll feeds, Google topics. That ate up 30-40 minutes before every single post. Now, once a week, I generate a bank of 50 ideas using ChatGPT-5 and just work my way down the list.
Here's the prompt I've saved as a template:
In 90 seconds, I get a month's worth of ideas. Half are predictable, but the other half are genuinely unexpected angles I'd never think of myself. If I use 30 out of 50, that alone pays for all my AI work, easily.
Step 2. Reels Scripts with Claude + Sora Preview
Reels are a whole other story. From one 90-minute podcast, I make five shorts: cutting out the strongest 30-60 second moments. That used to be a six-hour job for an editor. Now:
- I upload the podcast transcript to Claude 4.5 and ask it to pick out the five strongest moments, with explanations.
- For each, Claude writes a two-second hook (what the host says over the opening shot).
- If I don't have the right B-roll footage, I generate it with Sora 2: "B-roll cinematic shot of [podcast scene], 5 seconds, no audio."
- Then I load everything into editing software and assemble the shorts using the ready-made list of moments.
Assembling five shorts from one podcast now takes 1.5 hours instead of six. You can learn more about working with Sora and Veo in my Sora 2 vs. Veo 3 comparison.
Step 3. YouTube Thumbnails with FLUX 2 Pro
A YouTube thumbnail is half your CTR. I used to pay a designer 3000₽ per thumbnail, and we'd typically go through 3-4 iterations before finalizing it. That meant 9000-12000₽ and a week's wait for just one episode. For four episodes a month, that's 40,000₽ just for design.
Now I make covers myself with FLUX 2 Pro in 30 minutes. Here's the prompt formula for my style (dark background + guest's face + big headline):
I add the final headline text in Figma (FLUX messes up letters, as I wrote in the guide on 12 techniques). 30 minutes instead of a week, and 3000₽ in credits instead of 12000₽ per episode.
Step 4. Podcast Shorts with TTS + Veo
Some of my Telegram posts work great as Reels. I didn't make videos from them before — filming myself just didn't fit my schedule. Now I do it with TTS + video generation.
Here's the algorithm: I take the post text, cut it down to 100-150 words. Then I voice it with ElevenLabs TTS using a voice close to mine (I trained it on my timbre). At the same time, I generate visual B-roll with Veo 3 to match the text. Then I stitch it together on one video editing track.
One post → one Reel in 20 minutes. For 5 posts a week, that's 5 extra Reels — a 40% boost in Instagram reach compared to last year. More on working with TTS is in the article about AI podcasts.
Step 5. Adapting one piece of content for 3 platforms
The main secret of a solo blogger: don't create new content for every platform. One long piece should become Telegram posts, Instagram Reels, YouTube Shorts, and Pinterest pins.
I do this with Claude 4.5. One prompt:
Claude spits out the whole set in 2 minutes. Then I spend 30-40 minutes manually — final editing to match my voice. One podcast = 4 posts + 5 Reels + 3 articles + 8 quotes. Before AI, all this would take me a week. Now it's two days.
Tool Stack: What for What
| Task | Model | Cost |
|---|---|---|
| Ideas & Headlines | ChatGPT-5 | ~2-5 credits |
| Long-form Text, Scripts | Claude 4.5 Sonnet | ~5-10 credits |
| YouTube Thumbnails | FLUX 2 Pro | 8 credits |
| Post Illustrations | Gemini Image | ~5 credits |
| B-roll Video (up to 10s) | Sora 2 / Veo 3 | 15-25 credits |
| Voiceovers for Shorts | ElevenLabs TTS | ~80 credits/min |
| Podcast Transcriptions | Whisper | ~50 credits/hour |
Monthly Credit Spend
Here's my average monthly credit spend in Quantium, based on active use (4 episodes + 20 shorts + 20 posts + Pinterest):
| Activity | Credits |
|---|---|
| Ideas + Headlines + Text Editing | ~2500 |
| 4 YouTube Thumbnails + Variations | ~400 |
| 40 Post Illustrations | ~1200 |
| 15 Short Video Inserts | ~2000 |
| Voiceovers for 10 Short Videos | ~1500 |
| 4 Podcast Transcriptions | ~400 |
| Total | ~8000 |
This is fully covered by the Quantium Pro plan for 12,000 credits. Compare that to individual service subscriptions: ChatGPT Plus $20 + Midjourney $30 + ElevenLabs $22 + Claude Pro $20 + Sora $20 = $112/month. And that's without Veo, Gemini Image, and other models I need.
Why an Aggregator Beats Individual Subscriptions
As a solo blogger, you don't have a team to manage subscriptions. Each service needs its own card, its own interface, its own password, its own limit. You'll spend half an hour a month just trying to remember where you left off yesterday.
With Quantium, I get one Telegram bot, one credit balance, one monthly payment, and access to all 30+ models I mentioned. It's not just about saving money (though you do), it's about saving mental load — a solo creator's most valuable resource.
What Grew in Numbers Over the Year
Here are the metrics I track monthly:
- YouTube subscribers: 120k → 180k (+50%). The main factor: a steady 4 releases per month instead of 2, plus better covers.
- Telegram reach per post: 18k → 42k (+133%). Regularity (5 posts a week) helped, as did more precise headline work with ChatGPT-5.
- Instagram Reels reach: 35k → 110k (+214%). This was the strongest growth, thanks to shorts from podcasts and posts via TTS + Veo.
- Ad and sponsor revenue: grew 2.4x due to higher content volume and cleaner analytics across platforms.
- Hours worked per week: 60 → 28 (without losing quality). That means I had time for actual expertise — meetings, reading, research — not just production.
Key takeaway: AI didn't turn me into a media conglomerate. I'm still one author, one voice. But my productivity ceiling shifted from 2 releases a month to 4-5, and that changes the whole solo-blog economy.
5 Mistakes I Made My First Year
If you're bringing AI into your blog, don't repeat my mistakes.
- Trying to generate all content completely. For the first three months, I tried "ChatGPT writes — I publish." Subscribers noticed in a week; ER dropped 30%. Now AI just does drafts; final editing is always me.
- Using one model for everything. I stuck with ChatGPT for too long, which created a monotonous tone. Bringing in Claude for storytelling drastically changed the quality of story posts.
- Ignoring prompt engineering. I wrote vague prompts and got vague results. Now I've got 20 saved templates for different formats — that saved another 30% of my time.
- Too much AI visual content. When I made all YouTube covers through FLUX, they started looking identical. Now I shoot half the covers myself on my phone — that gives me the variety I need.
- Didn't track expenses. For the first two months, I didn't track how many credits went where. When I started, I saw 40% went to useless video generation experiments. Now I've got a budget by category.
Related materials: marketer case study, 12 FLUX 2 Pro techniques, TTS podcasts, Sora vs Veo.
Frequently Asked Questions
How many credits does an active blogger use per month?
An average solo blogger across YouTube, Telegram, and Instagram typically burns through 7,000-10,000 credits monthly. That's 40% on chat models for ideas and scripts, 35% for images for covers and posts, 15% on video generation (Sora, Veo), and 10% on TTS. A Quantium Pro subscription with 12,000 credits covers it with plenty to spare.
Can one person manage three platforms with AI?
Yep. The main idea is to shoot one long piece of content (a YouTube video or podcast), then use AI to adapt it for Reels, Shorts, Telegram posts, and pins. One hour of shooting can give you 1 video, 5 shorts, 8 posts, and 12 quote cards. Doing that alone without AI? Impossible.
Which AI tools does a blogger need first?
Minimal stack: ChatGPT-5 for ideas and headlines, Claude 4.5 for scripts and text adaptation, FLUX 2 Pro for YouTube covers and posts, Sora 2 for short video inserts, ElevenLabs TTS for voiceovers. With Quantium, you get all of them under one subscription, no jumping between services.
Why is an aggregator better than separate subscriptions?
A solo blogger just can't afford $20 for ChatGPT + $30 for Midjourney + $22 for ElevenLabs + $20 for Claude. That's $90+ every month. Quantium Pro gives you access to all those models, plus 25+ more AI tools, for less. And you get one interface in Telegram instead of juggling five.
Will subscribers notice I'm using AI?
If you use it smart, no. AI helps with drafts, ideas, and technical tasks, but your final content still comes through your voice, experience, and personality. Subscribers connect with the creator's voice, not the tool. The problem only hits when AI generates *all* the content — that's when a blog loses its identity.
Try Quantium for Free
Get 20 credits monthly on the free plan. 30+ AI tools in one Telegram bot.
Open Bot →Read Next
