Eastern European records: what FamilyLore reads

Eastern European handwriting from our shelf

A dotted underline marks a word FamilyLore was unsure of; [?] is its own note on a doubtful reading.

Matvei Lapatin to Grigorii Shelikhov · 24 Oct 1792

1. 5Милостивый Гдрь;
2. 6Григорьи[?] Ивановичь;
3. 7Имелъ честь получить отвас пи
4. 8Смо отъ числа сего мца чрезъ игнатья
5. 9петровича котелникова прикоторомъ
6. 10итоварной регистръ сценою Суммою на
7. 117256[?] р.[?] изначущимися понакладной фанту[?]

See the whole reading and its translation →

Taras Shevchenko to Varfolomii Shevchenko · 22 Apr 1860

1. 1[stamp: Ч. Г. З. МУЗЕЙ украин. древн. В.В. ТАРНОВСКАГО № 14 III]
2. 2№ 14.
3. 322. Березня
4. 41860.
5. 5Сёгодня [?] я получив твоє пись-
6. 6ма. Сёгодня у мене велике свято.
7. 7Сёгодня 22 березня 1838 року я

See the whole reading and its translation →

Also on the shelf: Frédéric Chopin’s 1831 letter in Polish cursive · Tekla Justyna Krzyżanowska’s 1782 baptism in Latin · Johann Kilian’s 1858 letter in Kurrent

How well it reads, measured

Measured on historical European handwriting. Tabular registers and columnar documents are not yet measured. How well FamilyLore reads →

Questions

Can’t ChatGPT do this?

A general chatbot can read some Eastern European hands, sometimes well, but it usually hands you clean text with no sign of which words it guessed, and its makers publish no figures for this script; FamilyLore marks the words it is unsure of and publishes how often it reads correctly, in the table above.

Which Eastern European scripts does FamilyLore support?

FamilyLore supports Russian cursive (both pre-1918 imperial orthography and Soviet eras), Ukrainian cursive, and German Kurrent and Sütterlin used throughout Central and Eastern European settlements.

Can FamilyLore read records written in Polish or Romanian?

Polish letters and Latin parish entries are read by FamilyLore (as shown on our shelf in the Chopin letter and Krzyżanowska baptism), but are not yet measured in the benchmark. Romanian handwriting has not been evaluated.

What about Russian-language records from Congress Poland?

You can upload them under the Russian choice (pre-1918 era). Note that columnar vital registers from Congress Poland are not yet measured.

What if it reads a word wrong?

Tap a word to see the other readings FamilyLore weighed, and correct it if you know better. The translation is redone from your correction. Most misread words carry a mark, but not every one, so a name or a date that matters is worth checking against the photo.

Who sees my pages?

Only you. A page becomes public only if you share it, and then only to people who have the link. Deleting your account deletes the photos you uploaded.

What FamilyLore doesn’t read well yet

• Parish registers, revision lists, and census books formatted in multi-column tables. Tabular forms are not yet measured.
• Church Slavonic script, and Romanian handwriting which has not been evaluated.
• Faded handwriting and damaged margins. Faded pages are measured on only a few pages so far.

From photo to text in three steps

1. 1

   Take a photo of the page

   Lay the sheet flat, in daylight and without flash. A phone photo of the whole page is enough.

2. 2

   Upload it

   “Russian or Ukrainian, Before 1918” is already chosen.

3. 3

   Read it

   You get the page as it was written, in modern spelling, and in English. Correct a word you know better yourself.