Docs

How this site builds itself out of a document tree.

Docs

Fourteen entries about one idea: that a website can be a document tree you edit in place, and everything else — routing, feeds, search, structured data — can be derived from it rather than configured.

Read them in order if you have twenty minutes. If you have two, read Start here and then click something.

guide

August 31, 2026

Start here

What you are looking at, and what to click first.

Read entry

Aug 29

Every page is a document

The tree is the site. Routing, feeds and search fall out of it.

Aug 27

The 23 page types

Editorial, collections, media, composition, visitor pages — and a globe.

Aug 25

Visitors are soft identities

Why registration must stay open, and why this demo lets you write.

Aug 23

Server-rendered, then live

Prose arrives in the HTML; the CRDT takes over on hydrate.

Aug 21

Meet the janitor

Someone is tidying up after you, by hand, character by character.

Aug 19

Run it yourself

Two containers, one command.

Aug 17

Editing is the studio

The owner surface: a bar, a right-click, and drag-and-drop.

Aug 15

Search, feeds and the sitemap

Three surfaces nobody maintains — all derived from the tree.

Aug 13

Presence is a person, not a tab

Who is here, where they are pointing, and why two tabs are one visitor.

Aug 11

Visitors write without write access

The guestbook and the hearts, on a site where visitors are read-only.

Aug 9

It works offline

Turn off the network and keep reading. Keep writing, too.

Aug 7

Making it look like yours

One grey, one accent, and a backdrop — from the settings, live.

Aug 5

A page can have pages

Depth is free. Any document can hold documents, and the URLs follow.