Home / Articles / Storyboards get their own site

Storyboards get their own site

The storyboards started in a book, moved to gen-z-developers.sgit.ai and then to pki.sgit.ai. Today they move into a site of their own, with the code that draws them on a second site, so any page can use it.

The storyboards on this site were not designed as a product. They were a way to write a book: before a chapter was written, it was a strip of panels with a punchline, an idea, the evidence and the people it was for. On 18 October 2026 gen-z-developers.sgit.ai drew the first of them from data, with no image model involved (Storyboards drawn by code). Two days later pki.sgit.ai vendored the same kit, added keys, vaults and QR codes to its props, and drew seven boards about passkeys and verification (pki.sgit.ai's storyboards).

Two sites carrying two copies of the same file is the moment a thing wants a home. The brief says so in its own words: the storyboards are "moving into its own world, which is here".

What moved, and what did not

BeforeNow
The storyboards16 on gen-z-developers.sgit.ai, 7 on pki.sgit.aiAll 23 here, imported byte for byte, each naming the commit it came from
The kit that draws themassets/storyboard.js copied into each site, one with a local patchOne kit, served by api.storyboards.sgit.ai, pinned by path and by hash
The propsgen-z's six; pki's twelve in a patch to the kitgen-z's six in the kit; pki's twelve as a prop pack any storyboard can load
The shapes a storyboard comes inA page, and its JSONA page, its JSON, a bundle for embedding, a markdown script, a semantic graph, a thumbnail, a link preview, and SVG, animated SVG, PNG and JPEG on demand

What did not move is authorship. A gen-z storyboard is still gen-z-developers.sgit.ai's file, checked there against the book's own words; this site shows it, links each board to the exact file and commit it came from, and records in content/series/<series>/series.json everything it adds (the shelves, the blurbs, how each character stands on the shelf). The origin sites keep their own copies for now. Pointing them at the API instead is the next step, and it is theirs to take.

Two repositories, one agent

The brief asked for something the newsroom does not yet do: the site and its API built side by side, by one agent, kept in step on every change. So this site does not merely link to the API, it depends on it in the way any other site would. Every page loads the kit from https://api.storyboards.sgit.ai/v0/v0.1/v0.1.0/, with an integrity attribute copied from that release's manifest, and the build fetches the API's own formats.js to write the markdown and the graph of each storyboard, checking each byte against the pinned hash before running it. If the API changed a published file, this site would stop loading it, by design.

What this release does not do

It does not let you make your own storyboard or your own cast yet; the brief puts that later, and says first to replicate what exists. It does not sell anything. The images it commits are the cover of each storyboard at two sizes; every other picture is drawn in your browser when you ask for it, because a picture made from data is a copy of the data and the data is what gets corrected. And it has not been read by the person who asked for it: the brief's second part is an agent's reading of a voice memo, and the corrections file is where it will be put right.