I Was the First One Who Needed It.

Late at night. I’ve just finished a last line — landing copy, the body of an onboarding email, a section of documentation I’ll let go tomorrow — and I try to re-read it.

I can’t.

Not the romantic exhaustion of the late-night founder. Something smaller, more embarrassing. I am too far in. My head is still thinking in Svelte, in n8n webhooks, in payloads and redirects. It cannot come back to the text the way someone would read it who opens the page for the first time. That jump — from the one who wrote to the one who reads — I am not making it. The muscle is not there.

That is where the idea of the four agents came from. A night like any other.

I am a solo founder. For months I have been building Obliqo: frontend, backend, automations, the Chrome extension that shipped recently, infrastructure self-hosted in Germany, payments. Hundreds of hours. This is not a heroic story. It is a fact.

An AI model gives you text that looks finished before the thinking behind it is.

The tool I am building exists for one specific reason: to slow down before publishing. To force yourself to read what you have written with different eyes, before a compliant AI model flatters it and lets it pass. Four voices with no stake in agreeing with me, that push back, that ask whether what I wrote actually holds.

I built it at the speed it exists against.

I was the first one who had to use it. And I needed it precisely because of how I was building it.

This is not the “dogfooding” of indie-hacker stories. That is the virtuous version: I used my own product, I liked it, I improved it. This is not that. This is dogfounding in the cruder sense: I needed a version of myself that had not just written the thing I had to re-read. A version of myself that had not been awake for eighteen hours. A version of myself with enough mental room to say to my face, “no — here you’re promising more than you can hold.”

There are two ways to tell this story, and I refuse both.

The first is the heroic story of the solo founder who worked late into the night, suffered, made it. Grind-porn. We know it. The hundreds of hours are not a trophy — they are the price of a choice I made, that I might make again, that I might not.

The second is the subtler version: confession-as-medal. The founder who admits his own exhaustion to turn it into a humility performance — a stoic pose, a calibrated display of self-awareness for the audience. That is grind-porn dressed as remorse. More dishonest than the first because it pretends to be its opposite. I do not fall for it. I try not to.

What is left, with both narratives stripped away, is something smaller and more uncomfortable. I needed my own product because I was working in conditions where I knew I would publish badly without it. Not virtue. Not discipline. Necessity under pressure. The tool came from there — not from brainstorming, not from customer interviews, not from a product roadmap. From a night when I could not re-read myself.

For someone who does not build software but writes things they publish under their own name — a post, a proposal, an important email, a PR description — this detail might matter. One person built this, and that same person tells you to your face what the tool does NOT do: does not rewrite, does not flatter, does not correct. This is not the pitch. It is that the syntax of the product is the same as the syntax of the hours it was made in. For some, perhaps, trust starts there.

The extension is live on the Chrome Web Store. It activates inside the tab where you are writing — Gmail, a PR description, the body of a post — and reads the draft before you send it. The webapp keeps living at obliqo.pyragogy.org. The extension is the same thing, brought to where the work happens.

A note: when you install, Chrome shows a warning, something like “extension not trusted.” It does not mean the extension is dangerous. It means I am a new developer and Google trusts those who have time behind them. Time I do not have yet. It dissolves on its own, month after month. Chrome is asking me to continue — annoying, but also fair: I built a tool on the trust you earn over time, and the browser is asking me to earn it. Fair enough.

I still do not know how the contradiction dissolves — building a tool against frenzy while I was inside the frenzy. Maybe it does not dissolve. Maybe the product, in the end, is the sediment of that contradiction — four voices doing for the reader what I could not do for myself, in time, alone. Maybe this is the limit of a single founder, and maybe it is also why the tool makes sense.

Now I will read this page with Obliqo. If it tells me not to publish, I do not publish.

If it tells me to publish anyway, it is probably lying to me. In that case, I decide.