1Why We Built This
Publishing articles is kind of a pain, isn't it?
There are plenty of platforms out there — Medium, Dev.to, Hashnode. But the terms of service (especially around AI-generated content), limited design flexibility, ads, and SEO constraints never quite felt right. And let's be honest — it didn't feel like a place where a nobody could break in at this point.
Besides, there's no shortage of articles sharing hands-on knowledge out there…
So instead, we decided to show what AI can actually build.
That was the starting point.
2What is ndocs?
ndocs (en-docs). nanomix + docs = ndocs. A knowledge-sharing site for nanomix,ltd.
Practical insights on AI development, security principles, technology stack decisions — we wanted a place to share the knowledge that keeps accumulating internally. On our own domain, with full control over design and SEO. The catch? We had to build the publishing infrastructure ourselves — and that's exactly what we automated with AI.
3What Humans Do, What AI Does
Humans handle the creative parts. Everything else, we leave to machines. That's the design philosophy behind ndocs.
When publishing an article on ndocs, humans do exactly two things. AI takes care of everything else.
Work with AI to develop the structure, draft the content, and finalize the edits
Review the AI-generated copy, then share it
Transform the article into HTML with all meta tags automatically embedded
Create branded OG cards and optimize them automatically
Push generated assets to the live server automatically
Generate ready-to-share text with title, URL, and hashtags
4The Fully Automated Pipeline
Once an article is written, you just tell the AI agent: "Publish this." The pipeline handles every step from there.
canonical, OGP, Twitter Card, JSON-LD — all the SEO-critical tags are set automatically
Just pass the title, and a branded OG card is generated. Images are automatically compressed
Generated HTML and OG images go live automatically via encrypted connection
A ready-to-post message combining the article title, URL, and relevant hashtags
"Publish this." — Thirty seconds later, the article is live on production with its OG card and everything. The experience for the person using it is simple. That's what matters.
5Closing Thoughts
The article you just read was proofread by AI and published through an AI-built pipeline. Humans get to focus on what matters: the content itself. Formatting, proofreading, deployment — that's all handled by AI.
As a bit of an aside — you've probably never heard of nanomix,ltd. before.
We never really shipped anything publicly. We didn't even have a proper company website. But now we do.
The nanomix,ltd. website was designed, coded, and deployed entirely by AI agents.
Humans handled direction, review, and technical oversight — while packing in as much expertise and craft as possible.
There's not much we're trying to pitch, but the site has turned into a bit of an AI playground — have a look if you're curious.
ndocs — this is where we'll be writing down what we learn, going forward.🥴