Free, permanent identifiers for published works and the products made from them. Built on ARK, so global uniqueness and resolution are handled by infrastructure that already exists rather than reinvented here.
It does not replace an ISBN if you want print distribution, bookstore ordering, or library sales. Those systems run on ISBN and will keep running on ISBN. If you publish direct or digital-only, this stands on its own; if you need the print supply chain, buy the ISBN and carry both. Anyone who tells you otherwise is going to cost you a distribution deal.
| POST /mint | Mint a Work or Product. Bearer token. Requires evidence the work exists. |
|---|---|
| GET /ark:/99999/… | Resolve. HTML or JSON by content negotiation. Tombstones return 410 with metadata. |
| GET /ark:/99999/…/history | The append-only event log. |
| GET /lookup?type=isbn13&value=… | Crosswalk from an existing identifier. |
| GET /onix/<blade>.xml | ONIX 3.0 Product record, ProductIDType 35. |
| GET /snapshot.ndjson | The whole public registry, one record per line. |
| GET /stats | Record and event counts. |
Every publisher gets their own shoulder, issued by software, with no application and no review — not by the ARK Alliance and not by us. Shoulders are opaque and assigned rather than chosen, so nobody can claim /penguin and turn this into a trademark court. Your name lives in metadata, where a name is a claim rather than a grant of ownership.
Minting has a volume ceiling that rises when you verify an email or a domain. That limits the faucet, never the door: no one here decides whether your book deserves an identifier. If you dislike the limits, run your own node — the code and the data are both open.