AI & Digital ToolsCommon Good, Economy & Energy

Thought #465: Transparency about financing, my own budget and running costs: Why I share my numbers in proportion – from 500 dollars of hard-learned lessons to a 130-euro setup

Published on

Look, when you build a project like uno.team the way I do – completely independently, with no investors, no public funding and no advertising behind you – at some point a very simple question comes up: how does this actually get financed day to day out of your own budget, and what does this whole engine room cost per month?

I want to be just as open with you about this as I am in my personal diary entries. Not with bare bank statements down to the cent, because that doesn’t help anyone online, but in honest proportions. Because especially when you have to get by on limited means of your own in everyday life and there’s no big company behind you, every euro counts twice.

At the start, working with the AI models, I first paid some real tuition – I learned the hard way. Before our current workflow was in place, I burned through around 500 US dollars in extra usage credits – simply because every deep code loop, every automated test run over more than a thousand files and every correction, billed call by call, cost money straight away. For my personal budget that was a real warning shot: it couldn’t go on like this.

So I had to change the cost structure radically. Today the entire development and quality assurance runs on a fixed, predictable setup: the central building block of our programming and checking pipeline in Claude Code is the Claude Max plan at 90 euros net (107.10 euros including VAT) per month, plus my regular Gemini subscription for drafts and fact-checks and a few euros for the domain, while hosting on Cloudflare Pages costs nothing in day-to-day operation. Instead of unpredictable 500-dollar spikes, the running tool costs now come to around 130 euros a month in total, VAT included – a frame I can carry myself, even out of a tight private budget.

And now just put these roughly 130 euros a month in proportion to what has been created here, technically and in content. We’re now talking about more than 1,580 files, four complete language versions in German, English, Russian and Ukrainian with around 168,000 translated words, and 53 automated test suites with almost 18,000 individual checks on every single release.

If you commission a platform like this the classic way in the German-speaking region – with software development, a multilingual architecture, SEO checks and specialist translation that is clean both legally and linguistically – even with freelance specialists you quickly end up at around 62,000 to 86,000 euros, and with a digital agency plus a translation office at around 89,000 to 124,000 euros.

For me, that’s exactly where the real message lies: using AI isn’t about replacing people at all. My voice, my life experience from the GDR, from Leipzig and from Ukraine, my stance and the final decision – no machine can take those off my hands. But this team of three – my human direction and two AI systems that check each other – for the first time puts the tools into the hands of a single person without much capital to build something at eye level that used to be possible only with a big budget.

How do you handle this upheaval in your own projects or at work: did you also pay your tuition first with new tools – and where does technology give you real freedom today that you otherwise couldn’t afford at all?

Show fact-check & sources
Something to think about

What do you take away from this thought?

Three ways from here