Every lookup provider in Shelvd: what it holds, what it's good at, and where it falls short. Consider this your field guide to the world's library catalogs.
The Idea
No single library has every book. Not the Library of Congress, not the British Library, not that suspiciously well-stocked secondhand shop in Hay-on-Wye. But between them, the world's libraries hold more bibliographic data than any collector could generate in a lifetime.
Shelvd connects to 30 providers across 30 countries on 4 continents. One search, thirty libraries. Here's what each one brings to the table.
General & Multi-National
Open Library (🌐)
What: Community-maintained catalog with 20M+ editions. Part of the Internet Archive. Best for: Modern books, English-language titles, cover images. Quirks: Data quality varies wildly — it's crowd-sourced. Some records are pristine; others look like they were entered during a power cut. But the coverage is unmatched for post-1900 English titles.
Google Books (🌐)
What: Google's book database. You know Google. Best for: Quick ISBN lookups, cover images, descriptions. Quirks: Rich but shallow. Good at finding books, less good at bibliographic precision. Publisher names are sometimes the imprint, sometimes the parent company, sometimes a mystery. But for getting a cover image and basic metadata, it's fast and reliable.
WorldCat / OCLC (🌐)
What: The world's largest library catalog — 500M+ records from 100,000+ libraries. Best for: Finding the most widely-held edition, OCLC numbers, broad coverage. Quirks: Uses the Classify API, which returns aggregate data. Good for identifying which edition you have, less good for granular bibliographic detail.
National Libraries — Western Europe
Library of Congress (🇺🇸)
What: The largest library in the world. 170M+ items. Best for: LCCN, LCC, DDC numbers. Authoritative records for English-language titles. Quirks: SRU interface with MARCXML. Some older records are skeletal, but modern cataloging is exceptional. If the LoC has cataloged your book, the data is likely the best you'll find.
British Library (🇬🇧)
What: The UK's national library. 170M+ items including the Magna Carta. Best for: UK publications (legal deposit), historical British publishing, ESTC overlap. Quirks: SRU/MARCXML. Recently migrated to a new discovery system. Excellent for anything published in Britain.
What: France's national library. 40M+ items. That building on the Seine that looks like four open books. Best for: French-language publications, UNIMARC data, BnF identifiers. Quirks: Uses UNIMARC (fields 200, 210, etc.) instead of MARC21. CQL queries need specific relation operators. Shelvd handles the translation.
SUDOC (🇫🇷)
What: Union catalog of French university libraries. 15M+ records. Best for: Academic French titles, theses, dissertations. Quirks: Also UNIMARC. Field 214 for recent publications (replaces 210). Good complement to BnF.
DNB — Deutsche Nationalbibliothek (🇩🇪)
What: Germany's national library. Everything published in German since 1913. Best for: German-language publications, GND authority data. Quirks: SRU/MARCXML. Excellent authority control. NSB/NSE markup (non-sorting characters) needs cleanup — Shelvd strips these automatically.
K10plus / GBV (🇩🇪)
What: Union catalog of German and Dutch research libraries. 200M+ titles. Best for: German academic titles, broad European coverage. Quirks: PICA-based system. Uses the same SRU endpoint as CERL HPB but with different indexes.
KB — Koninklijke Bibliotheek (🇳🇱)
What: The Dutch national library. Best for: Dutch publications, historical Dutch printing. Quirks: Custom Dublin Core format via SRU. Good for anything published in the Netherlands.
KBR (🇧🇪)
What: Belgium's Royal Library. Everything published in Belgium. Best for: Belgian publications, Plantin-Moretus holdings, Flemish and Walloon printing. Quirks: SRU/MARCXML. Because Belgium has three official languages, records can be in Dutch, French, or German.
Unicat (🇧🇪)
What: Belgian union catalog — all academic and research libraries. Best for: Belgian academic holdings, theses. Quirks: Broader than KBR alone. Good for finding which Belgian library holds a specific title.
BNE — Biblioteca Nacional de España (🇪🇸)
What: Spain's national library. Best for: Spanish-language publications, Latin American imprints. Quirks: SRU/MARCXML. Good for Iberian publishing history.
Swisscovery / SLSP (🇨🇭)
What: Swiss union catalog — 475 libraries. Best for: Swiss publications, multilingual records (DE/FR/IT). Quirks: Alma-based. Switzerland catalogues in three languages, so you might find the same book with German, French, and Italian subject headings.
National Libraries — Nordic & Beyond
BIBSYS / Oria (🇳🇴)
What: Norwegian union catalog — all academic and research libraries. Best for: Norwegian publications, Scandinavian titles. Quirks: Alma-based SRU. Good for Nordic academic coverage.
Libris (🇸🇪)
What: Swedish union catalog — 400+ libraries. Best for: Swedish publications, Nordic literature. Quirks: Custom XSearch API. Good for Scandinavian coverage alongside BIBSYS and Finna.
Finna (🇫🇮)
What: Finnish aggregator — museums, libraries, archives. Open data. Best for: Finnish publications, Nordic titles, open API. Quirks: JSON API. One of the most developer-friendly library APIs. Finnish libraries take open data seriously.
DanBib / bibliotek.dk (🇩🇰)
What: Danish union catalog. 14M+ records.
Best for: Danish publications, Scandinavian coverage.
Quirks: OpenSearch API with DKABM/Dublin Core XML. CQL queries use dkcclterm indexes. Year search uses dkcclterm.år — yes, with the Danish å.
ÖNB — Österreichische Nationalbibliothek (🇦🇹)
What: Austria's national library. One of the oldest in the world. Best for: Austrian publications, Habsburg-era printing. Quirks: Alma-based SRU. Good for Central European bibliography.
COBISS (🇸🇮🇷🇸🇧🇬🇲🇰🇧🇦🇲🇪🇦🇱🇽🇰)
What: Co-operative Online Bibliographic System and Services — 10M+ records across 780+ libraries in 8 countries: Slovenia, Serbia, Bulgaria, North Macedonia, Bosnia-Herzegovina, Montenegro, Albania, and Kosovo.
Best for: Southeastern European publications. If it was published in the Western Balkans, it's probably here.
Quirks: No public API — uses the legacy COBISS+ interface which returns server-rendered HTML. Expert search with prefixes (BN= for ISBN, TI= for title, AU= for author). Searches Slovenia first (largest database) then cascades through other national systems. Developed by IZUM in Maribor since the 1980s — one of the longest-running cooperative library networks in Europe.
BnL — Bibliothèque nationale du Luxembourg (🇱🇺)
What: Luxembourg's national library and union catalog of 80+ libraries in the bibnet.lu network. 1.8M+ printed items. Legal deposit library for Luxembourg. Best for: Luxembourg publications and Luxemburgensia (works by or about Luxembourg, wherever published). Quirks: Runs on Ex Libris Alma — same SRU pattern as the Austrian National Library. Opened a stunning new building in Kirchberg in 2019. Also runs the Bicherbus, a mobile library serving 81 villages by refurbished bus.
Biblioteka Narodowa (🇵🇱)
What: Poland's national library — 4.7M+ bibliographic records, the largest bibliographic database in the country. Compiles the Polish national bibliography. Best for: Polish publications, Polish-language books, and works about Poland published abroad. Quirks: Excellent free REST API at data.bn.org.pl — returns JSON with full MARC21 records embedded. No authentication needed, unlimited access. Search by ISBN, title, author, publisher, year, material type. Also hosts Polona (polona.pl), a digital library of Polish cultural heritage.
South America
BN Brasil — Biblioteca Nacional (🇧🇷)
What: Founded in 1810 when the Portuguese royal family fled to Rio de Janeiro with their library. Today holds ~9M items — one of the largest libraries in Latin America, recognized by UNESCO. Best for: Brazilian publications. The CBL (Câmara Brasileira do Livro) is the official ISBN registry for Brazil — every Brazilian ISBN since 2020 is registered here. Quirks: Uses BrasilAPI, a free open-source API that aggregates data from CBL and Mercado Editorial. ISBN lookup only — no title/author search. Response includes title, authors, publisher, year, page count, subjects, synopsis, and sometimes cover images. Brazilian ISBNs use prefix 65- or 85-. Authors often returned in ALL CAPS (normalized automatically).
Middle East
NLI — National Library of Israel (🇮🇱)
What: Israel's national library, founded in 1892 as a small collection in Jerusalem. Today holds ~5M items including the world's largest collection of Hebraica and Judaica. Best for: Hebrew-language publications, Judaica, Middle Eastern studies. Quirks: Migrated to Ex Libris Alma in 2020, so SRU works the same as ONB, BnL, SLSP, and BIBSYS. Records in both Hebrew and Latin scripts. Discovery interface at primo.nli.org.il.
Asia-Pacific
NDL — National Diet Library (🇯🇵)
What: Japan's national library. The largest in Asia. Best for: Japanese publications, CJK metadata. Quirks: OpenSearch API returning RSS/Dublin Core XML. Shelvd has a custom parser. Excellent for Japanese editions.
Trove / NLA (🇦🇺)
What: Australia's national discovery service — libraries, newspapers, archives. Best for: Australian publications, Oceania coverage. Quirks: JSON API requiring an API key. Currently pending key approval — the provider is registered but temporarily disabled.
Specialist Catalogs
OPAC SBN (🇮🇹)
What: Italy's national union catalog — the Servizio Bibliotecario Nazionale. Best for: Italian publications, Italian printing history. Quirks: Custom JSON API with its own query language. Italian libraries have their own way of doing things.
CERL HPB — Heritage of the Printed Book (🇪🇺)
What: 6M+ records of European rare books (1455–1830). Run by the Consortium of European Research Libraries.
Best for: Incunabula, early printed books, provenance research. If your book was printed before 1830, start here.
Quirks: SRU with PICA indexes (pica.tit, pica.per, pica.yop). Returns MARCXML with rich provenance data: author life dates, printers, former owners, physical dimensions, binding notes. The pica.yop index only supports exact year — no ranges.
HathiTrust (🇺🇸)
What: 13M+ digitized volumes from 200+ research libraries. Best for: Checking if a digital version exists, finding which libraries hold a copy. Quirks: ISBN/OCLC/LCCN lookup only — no title or author search. This is an identifier-based API, not a search engine. But it tells you things no one else does: which university holds the book, whether a full-view digital copy exists, and the rights status.
Europeana (🇪🇺)
What: 200M+ records from 4,000+ cultural institutions across 49 European countries. The mother of all European aggregators.
Best for: Pan-European fallback when national libraries come up empty. Also covers manuscripts, maps, prints, and other cultural objects beyond books.
Quirks: REST API with JSON responses. Searches across the Europeana Data Model (EDM), which means fields are language-tagged and can vary wildly between contributing institutions. ISBNs live in dcIdentifier rather than a dedicated field, so results need filtering. Filtered to TYPE:TEXT to keep it book-focused. Think of it as WorldCat's European cousin — breadth over depth.
BNP / PORBASE (🇵🇹)
What: Union catalog of Portuguese libraries. PORBASE aggregates records from the Biblioteca Nacional de Portugal and dozens of academic and public libraries across the country. Best for: Portuguese-language editions, books published in Portugal, and Lusophone literature in general. Quirks: Not SRU — uses a URN HTTP service that returns MODS XML for a given ISBN. Dead simple: one ISBN in, one record out. No multi-field search (title, author) — ISBN only. But the MODS response is clean and well-structured: title, authors with roles, publisher, place, year, extent, language, edition, even UDC classification codes.
Choosing the Right Provider
- Modern books with ISBN: Open Library, Google Books, your national library
- Academic titles: WorldCat, SUDOC, BIBSYS
- European rare books (pre-1830): CERL HPB, then national libraries
- Digital availability: HathiTrust
- Pan-European fallback: Europeana
- Regional publishing: Use the national library of the country of publication
- Portuguese: BNP / PORBASE
- Scandinavian: Libris (SE), BIBSYS (NO), Finna (FI), DanBib (DK)
You can enable and prioritize providers in Settings → Book Lookup. There's no wrong order, but starting with your country's national library and Open Library covers most cases.
See also: Library Lookup · Configuring Providers · Enriching a Record