Reintroducing Intracia

Three launches on one day, and I've been looking forward to this one for a while. Most of the last year went into the parts you'll never notice: the schemas, the diffs, the hundred small decisions that add up to an editor which feels intuitive to use.

The site

You're reading it: Markdown in a Git repository, built once, served from a CDN. No database, no application server, no admin panel sitting on the public internet waiting to be found.

That last part matters more than it sounds. Most website breaches start with something that executes — a plugin, an outdated CMS, a login form. A static site has almost none of that to attack. There's nothing to inject into, and nothing to escalate through.

It's also close to free to run, and the price doesn't move when the traffic does. Serving a file to 10,000 people costs roughly what serving it to 10 does, because a CDN is doing the work rather than a server you're paying for by the hour. A front-page spike is a bandwidth line item, not an outage and a 3am phone call.

And it's fast, which is the part everyone notices. The page is already built and sitting near the visitor when they ask for it.

The catch has always been the editing. More on that below.

What we take on

Intracia has been running since 2020, mostly on work that arrived through people we'd worked with before. That's a pleasant way to run a business and a poor way to be findable, so the site now introduces us properly.

I've spent four decades on the tools other developers depend on — LINQ to SQL and Entity Framework at Microsoft, the Atom editor at GitHub, SDKs and developer experience at Auth0, the official MongoDB provider for EF Core. Before any of that, the first internet banking system built on Temenos Globus. Kerry leads marketing and design here and runs MKG Marketing alongside it, so between us a project doesn't have its interesting half handed to somebody else.

That's what shapes the four things we take on:

• Software development — applications and APIs, designed by someone who has had to operate them afterwards.
• Web applications & sites — customer-facing applications, product interfaces, marketing sites and documentation, built end to end.
• Data extraction & recovery — getting data out of systems that would rather keep it. Undocumented formats, dead platforms, failing media, legacy databases.
• Developer & AI tooling — libraries, SDKs and internal tooling, plus the MCP servers and evaluation harnesses that decide whether an AI feature survives contact with production.

The list is short on purpose. It's the set we can start on without a learning curve you'd be funding.

Intracia

This is the part I've been wanting to talk about for a while.

If you run a content-heavy site there are two well-worn options, and both cost you something. Keep the content as Markdown in a repository and you get portability, review, history and builds you can reproduce — at the price of asking writers to use a code editor, hand-write frontmatter, and check a dashboard somewhere else to find out whether anything shipped. Or move to a hosted CMS and get a good editor immediately, in exchange for your content living in somebody else's database.

Intracia is an attempt to stop that being a choice. It puts a structured editing workspace over the content already in your repository. Editors get real fields instead of raw Markdown. You define the shape once and everyone writes to it. Nothing moves: the files stay in your repository, the media stays in your bucket, the site builds and deploys where it already did.

We provide the workspace, not the custody. If we vanished tomorrow your site would carry on serving, because the only thing we were ever holding was the editing interface.

It sits on top of what you already run — Nuxt Content, Astro, Eleventy, Jekyll, Next.js, anything that reads Markdown, YAML or JSON from a repository — and deploys through the pipeline you already have. GitHub today, GitLab in development.

One caveat, and it's a real one. It's early access. Features will change, there's no availability commitment yet, and access goes out one site at a time so we can look at each properly. If you need guarantees rather than good intentions, this isn't the moment to put something critical on it.

If it sounds useful, join the waitlist. Tell us what you're running and we'll tell you quickly whether it fits.

What comes next

We'll keep writing about building and running sites on a static stack — what it costs, what breaks and what's worth it — the lessons of taking a small SaaS from idea to early access, and the features of Intracia as they ship.

Browse the blog, or follow along via RSS or Atom.