Early-modern Europe, the Latin liturgy, and three new ways to search
7 November 2026 · Nabu news
Four November days of development bring the library to 2,854,118 documents / 128,499,000 passages across 327 live sources (census 7 November 2026), with the dictionary layer at 5,281,593 entries across 956 shelves.
Early-modern Europe arrives as a wave. The London book trade’s own ledger — the Stationers’ Register, 1557–1640 — enters as 10,453 dated entries, beside the Statutes of the Realm (661 acts of the English parliament to 1713), the Evans-TCP and ECCO-TCP transcriptions of early American and eighteenth-century English printing, 3,067 mazarinades (the pamphlet war of the Fronde, 1648–1653), the Socinian correspondence network (1,869 letters, each queryable from both its writing and its receiving place), the 16th-century Slovene Protestant prints of Korpus 16, and Noscemus’s neo-Latin scientific corpus (1,038 works / 555,797 passages). With them lands LEME — the Lexicons of Early Modern English — as 311 period dictionaries totalling 943,211 entries: the words of the age, defined by the age.
The Latin liturgy, sung. Six liturgy and chant sources went live in one wave: the Cantus Database (449,132 chants, exact on the database’s own advertised total), GregoBase’s engraved Gregorian scores, the Divinum Officium Tridentine office, the Greek Orthodox liturgical corpus, Church Slavonic service texts from Ponomar, and the Analecta Hymnica. A new chant-identity layer ties the same chant across witnesses — 71,032 edges at first build.
East and Inner Asia, broader. Seven sources in one window: Armenian epigraphy, the Jesuit mission press’s romanized Japanese (the Kirishitan prints), the Tangut rhyme dictionary of the Digital Wenhai, the Qing Pentaglot with its five-script rows, and the Việt–Tai lexica with their Han-script layers. The runic shelf gained its catalogue layer too: dating, findspots and inscription sigla harvested from RuneS-DB’s own records, with 370 translation siblings.
Three new ways to search. The library’s annotation layers —
standardized and regularized spellings, modern readings, margin
notes, editors’ and translators’ notes — are now a search surface of
their own: 2,336,455 passages across 31 sources answer
search --annotations (built 7 November 2026), so a modern-spelling
query reaches the period spelling that carries it, and a margin note
finds its page. Scholars’ token glosses became content too:
search --gloss-lines queries the glossing line of 320,055 passages
across five annotated corpora — English over the Hebrew Bible and the
Achemenet Akkadian, German over the Egyptian and Hittite corpora —
and returns the original-language passage, so “king” finds מֶלֶךְ and
König finds nswt. And upstream morphology tags across five more
tagsets now decode to readable grammatical features at display time,
with a word-grain gloss filter (--gloss) joining the lemma search.
The works taxonomy grew five ruled heads — theology with its
treatise, apologetics and polemic subdivisions, oratory, catechesis,
paratext, hagiography — toward a 28-head classification of what each
work is.
The demo keeps earning trust. The MCP demo server now ships with a 36-task evaluation pack exercising all 14 tools against the live image — a machine-checkable answer to “does it actually work” that runs before every release. Installing it is one click from the project page: VS Code and Cursor install badges, a Docker one-liner, and free-tier deploy buttons, with the image version now locked to the library’s own release line.
As always, the numbers above are the generated census of the live catalog, not hand-kept claims.