What I Already Own

Sources and method

The What I Already Own catalogue is built from open databases that allow commercial reuse. Here is where the data comes from, under which licence, how often it is rebuilt, how it is cleaned and counted, and what is left out.

Updated September 29, 2026

Where does the data come from?

Each source was chosen after reading its licence. Dates are those of the extracts in service.

Data from these sources has been reformatted, deduplicated and merged by What I Already Own, and may therefore differ from its original publication.

How often is the catalogue rebuilt?

  • Every month, the catalogue is rebuilt from scratch as soon as Discogs publishes its monthly dump: MusicBrainz, Discogs, Open Library, the national libraries and Wikidata are downloaded again, and the new catalogue is built next to the one in service.
  • It replaces the old one only after validation: the number of editions must not drop by more than 2%, and 500 reference barcodes must still find their edition.
  • Every week, new books from the national libraries of Germany, France, Spain and Japan are added without waiting for the rebuild.
  • A scanned book the catalogue does not know is looked up by its ISBN in national libraries and Open Library within fifteen minutes, at most one request per second.

How is the same edition recognised in two sources?

By its barcode first, then by label, catalogue number and country. Otherwise the two records stay separate: nothing is merged on the title alone. Each record keeps its source identifiers (MBID, Discogs release ID, ISBN).

How are albums, books and volumes counted?

The numbers announced at the top of discographies, bibliographies and reading orders are those of the lists shown. Whatever is set aside is counted below each list, never silently deleted.

  • Studio albums: live albums, compilations, remixes and other MusicBrainz secondary types are listed separately. The type comes from MusicBrainz as soon as one of the artist's works is there.
  • Soundtracks count as studio albums, unless recorded live.
  • A record on which the artist is only a sideman (neither first credited nor one of the two names of a duo) is removed from their discography, unless such records exceed 30% of theirs.
  • A book counts once, translations and reissues together. Collections and omnibus editions are removed, as are works that Open Library attributes to another author. A work known to Wikidata is never removed.
  • An isolated year more than ten years before the author's debut is replaced by that of the next edition. In a series, dates that break the order of volumes are removed.
  • A reading order is published only for a series numbered 1 to N with no gap, the order coming from Wikidata. A volume attested by Wikidata with no known edition is still shown.
  • A list stops at 150 lines, albums kept first; the number announced is the number shown.

Which pages are published?

A page is offered to search engines only if it carries a useful list: at least 10 dated works and 100 editions for an artist or an author, a complete series for a reading order. Names that pile up thousands of editions with a single publisher (notebooks, generated books) are left out.

What is left out

  • The Discogs API: its terms forbid commercial use. Only its monthly dumps, released under CC0, are used.
  • The MusicBrainz web service, whose live commercial use is paid: its CC0 dumps are hosted on our servers.
  • The Google Books API, Amazon, WorldCat and booksellers' sites: their terms forbid keeping their data.
  • No third-party API is called during a scan: everything is served from our own servers.

What about sleeves and covers?

They belong to their publishers and labels. They are shown to identify an edition, at reduced size, copied to our servers rather than linked from the source. Photos taken by users stay private by default.

What about items missing from the catalogue?

A user who scans an unknown item can create its record. It is accepted when a library confirms the same ISBN and a compatible title, otherwise reviewed by hand. Once validated, it benefits everyone, released under CC0.

Request a correction or a removal

Rights holders who want an image or a record removed, or anyone who spots an error, can write to: [email protected]

About What I Already Own