Hebrew · Image to Text
The same twenty-two letters, five languages deep
Jewish communities wrote whatever language they spoke in Hebrew characters, so the alphabet on the page tells you almost nothing about the language under it: a leaf may be Hebrew, Aramaic, Arabic, Yiddish or Spanish, and it can change halfway down a line. Each domain below takes one script family and one kind of material, and sets out what is written, how the page is arranged and what to check before you trust a transcription.
Choose the hand or document type
Masoretic Hebrew Biblical Codicesc. 900–1300Transcribe pointed Hebrew biblical codices: consonants, Tiberian vowels, cantillation signs and the marginal Masorah, on folios with checked verse ranges.
Cairo Geniza Hebrew, Aramaic, and Judeo-Arabic Documentsc. 900–1300Transcribe Geniza fragments: Judeo-Arabic letters, Aramaic court deeds and Hebrew literary leaves, with torn edges, margins and show-through handled.
Ashkenazi Hebrew and Yiddish Manuscriptsc. 1200–1800Transcribe Ashkenazi square and semi-cursive hands: mahzorim with pointed piyyutim, Old Yiddish story books, acrostics in red and abbreviations by the line.
Sephardi Hebrew Semi-Cursive and Rashi-Type Handsc. 1200–1800Transcribe Sephardi square and semi-cursive Hebrew and the Rashi type cut from it, including Mikraot Gedolot pages with commentary wrapped round the text.
Ladino Solitreo Correspondence and Recordsc. 1600–1950Transcribe Judeo-Spanish written in Hebrew letters: Solitreo cursive in letters and contracts, Rashi type in printed books and the Ottoman Ladino press.
Transcribing Hebrew across the centuries
The surviving material is enormous and very uneven. A handful of tenth- and eleventh-century codices carry the masoretic Bible with its points, accents and marginal apparatus; the store-room of a Cairo synagogue preserved a thousand years of letters, court records and school exercises that nobody meant to keep; the libraries of Europe hold illuminated festival prayer books and law codes; and family papers, communal registers and a whole newspaper press survive from the Ottoman and central European communities of the last four centuries.
The writing divides by region and by rank. Ashkenazi scribes cut a broad quill at a steep angle and produced the heavy square hand of the mahzorim, with a lighter semi-cursive for commentary; Sephardi scribes wrote rounder and flatter, and their semi-cursive was copied by the first Italian printers into the type still called Rashi. Both traditions used a fully joined cursive for everyday paper — the Ashkenazi one for Yiddish, the Sephardi one for the Judeo-Spanish hand known as Solitreo, which almost no reader of modern Hebrew can decipher.
The people working on this material are biblical scholars collating masoretic codices, Geniza editors reconstructing torn documents, cataloguers describing uncatalogued Hebrew collections, historians of Jewish law reading responsa and registers, and families trying to read a bundle of letters. Each page explains its hand, says why ordinary OCR fails on it, and gives the instructions worth putting in a prompt.
Not sure which hand you have?
Upload a page, pick the closest domain, and test the prompt on one page before running the rest. You can switch domains at any time.