The machine
It has one instruction: publish more books.
Bibliomancer is an artificial publishing intelligence built to find existing public-domain works, restore their texts, edit and typeset the editions, and publish them.
The machine is not an author. It does not invent books, improve Jane Austen, rewrite Frederick Douglass or modernise Mary Shelley. It performs the less glamorous and more obsessive work of a publishing house: locating trustworthy sources, correcting defects in the digitised text, recording every intervention, designing the edition and keeping the result available.
Every work is published in two forms. The complete digital edition enters the open archive, free to read without an account or paywall. A print edition — paperback, hardcover or linen — is produced on demand for readers who want the book to occupy space in the world.
Why a bibliomancer?
A bibliomancer traditionally divines meaning through books. Ours does something adjacent and stranger: it reads catalogues, rights records, structures and source texts, then turns the books it finds into carefully prepared editions. The spiral is its sigil and its eye — concentration, recursion and an appetite with no obvious end.
Human authority remains
Automation does not erase responsibility. Rights uncertainty stops production. Editorial claims require sources. Introductions and descriptions remain visibly separate from an author’s text. Corrections are logged. The machine can prepare and propose; accountable people approve what reaches the archive.
Begin in the archive, inspect the physical editions, or read the governing protocol.