Studio·July 2026

Building the AC Codex: Why an Editing Studio Made a Museum for a 20-Year Story

There's a story that has been running for twenty years across fourteen mainline games, a stack of side entries, a film nobody remembers and a modern-day conspiracy that most players skip the cutscenes for. I've followed it since I was a teenager. And for twenty years, the only way to actually see that story — all of it, in order, with the connections visible — was to open six wiki tabs and pray none of them spoiled an ending.

At some point last year I stopped waiting for someone to fix that and built it myself.

It's called the AC Codex — an interactive reference for the whole Assassin's Creed saga. Twenty-two entries, every era from ancient Greece to modern-day London, one canonical dataset underneath, and seven different ways of looking at it. It lives in our web portfolio now, and I want to write about building it, because the process ended up teaching me more about our day job — editing — than most editing projects do.

That sounds like a stretch. Stay with me.

The problem was never information. It was arrangement.

Everything the Codex contains already existed somewhere. Wikis have every date, every character, every artifact catalogued to exhaustion. If you want to know what happened in a game, you can find eleven thousand words on it in thirty seconds.

And that's exactly the problem. Eleven thousand words is not an answer, it's a warehouse. Nobody experiences a story as a warehouse. You experience it as a shape — this era led to that one, this character echoes that one, this game sits inside a trilogy the way a verse sits inside a song. The shape is the thing no wiki gives you, because wikis are built for completeness, and shape is built from omission.

So the design rule for the Codex, from day one, was museum-minimal: white space, one accent colour, and the confidence to show less. A museum doesn't put every artifact it owns on the floor. It picks, it sequences, it lights. The archive stays in the basement.

If you've read anything else on this journal, you already know why that rule felt familiar. It's the edit. It's exactly the edit. Every week we take two hours of a founder's footage and make ninety seconds that feel complete. The skill isn't collecting — it's deciding what the audience never sees.

One dataset, seven lenses

The technical idea that makes the Codex work is almost embarrassingly simple: there is exactly one source of truth — a dataset describing every game, era, connection, character and beat — and every page in the site is just a different lens pointed at it.

The main view is a chronological era tree, drawn as a living dendrogram with a faint Animus-style helix drifting behind it. Branches are eras; leaves are games; arcs overlay the connections between entries that share characters or artifacts. Toggle the side entries and the tree grows new twigs. It's the view I wanted for twenty years: the whole saga as one organism.

Then the same data, six more ways. A dual-track timeline that scrubs in-universe years against release years — one axis covers 2,500 years, the other covers twenty, and seeing them locked together is quietly hilarious. A world map with one red dot per setting. A poster gallery. A compare view that puts two to four games side by side in pure typography. Six curated story threads that follow a single idea — a bloodline, an artifact — across a dozen games. And a play-order guide that doesn't just recommend an order but argues with itself, because any fan will tell you there is no correct answer, only trade-offs.

Pick a game anywhere and it stays selected everywhere — the selection lives in the URL, so switching lenses never loses your place, and sharing a link shares your exact view. Press ⌘K and search the whole saga. Press 1 through 7 and jump between views. Small things. Small things are the job.

The spoiler problem, which is really a respect problem

My favourite feature is the least technical one. Every plot beat in the dataset carries a spoiler flag, and the site ships with spoiler mode on: ending material is blurred until you deliberately click it, and your choice persists. A new player and a twenty-year veteran can read the same page and have different pages.

I think about this more than I should, because it's a content problem disguised as a UI problem. Every piece of content has an audience that knows everything and an audience that knows nothing, and most content picks one and abandons the other. The blur is just a way of refusing to choose. Write for the veteran, protect the newcomer, let one click separate them.

There's a version of that in client work too. The founder watching their own video knows the whole story; the stranger scrolling past knows nothing. Every edit has to hold both. We just never had a blur tool for it.

How it's actually built

For the people who came here for the stack: Next.js 15 with the App Router, exported fully static. TypeScript everywhere. Tailwind for styling, Framer Motion for the transitions, and the d3 family — hierarchy, scale, shape, zoom — doing the heavy lifting on the tree. There is no backend and no database. The whole thing compiles to a folder of files and deploys to any static host in seconds.

The part I'm actually proud of is the pipeline, not the pixels. Adding a game to the Codex is two steps: append one record to one data file, and run one script that fetches the artwork, converts it to WebP at two resolutions, generates blur placeholders and writes the attribution manifest that feeds the credits page. Then a validation script checks the whole dataset — every connection points at a real entry, every era contains its games, every image exists — and fails loudly if anything is off.

That discipline — one source of truth, everything derived, a gate that refuses broken data — came straight from running a studio. Systems are the only reason four people can ship a month of content without dropping anything. The Codex is what happens when the same instinct is pointed at a hobby.

What it taught us

Somewhere in the middle of building the compare view, I realised I was making the same decisions I make in a colour-grading session. Which differences matter? Which numbers are noise? What does the eye need first? A comparison table and a ninety-second edit are the same object at different resolutions: both are arguments about what deserves attention.

So when clients ask why a video editing studio builds websites — and they do ask — this is the honest answer. We don't see a boundary. A story spread across twenty years of games and a story spread across two hours of interview footage have the same enemy: everything, presented equally, is nothing. The job is arrangement. The medium is negotiable.

The Codex is live in our web portfolio — the tree, the threads, the spoiler blur, all of it. If you love the series, start with the Threads view; if you've never played a minute, start with the play-order guide and let it argue at you. And if you're a founder wondering whether the people who built that can make your story make sense in ninety seconds — that is, genuinely, the same muscle.