Built with the book-data ecosystem

Kobbie does not split the pool evenly or by how many API requests it makes. Kobbie caches heavily on purpose, so counting requests would punish the sources whose data is good enough to be worth remembering.

Instead, Kobbie records which sources contributed the evidence behind each accepted piece of book knowledge: what a work is, who wrote it, which series it belongs to and where in that series it sits. Each quarter, Kobbie counts the distinct pieces of knowledge it actually used or served and divides the pool according to each source's measured share of that knowledge.

A book viewed a thousand times counts once, the same as a book nobody opened twice. If two sources contributed to the same fact, they each get half of it.

If a source helped establish a fact years ago and Kobbie has been reusing that fact ever since without contacting anyone, the source keeps its recognised contribution. Better caching and fairer support are meant to reinforce each other, not compete.

This is a voluntary contribution from Kobbie. It is not charity, and it is not payment for API access. It does not replace or extend any licence, agreement or permission Kobbie needs to use a source, and it does not grant Kobbie special treatment.

Naming a source here does not mean that source endorses Kobbie or is partnered with it. Sources currently eligible are Hardcover, Open Library, Wikidata and Wikipedia. Commercial paid services Kobbie buys from are part of Kobbie's costs and are not in this pool.

Of the $14.99 annual Plus price, $3.00 goes to this pool. The rest is gross revenue that still has to cover development, infrastructure, AI and provider costs, storage, payment fees, taxes and support. It is not profit.

Our first contribution period (2026-Q3) is currently accumulating. Figures appear here once a quarter closes. Nothing is published before it is real.