Genealogy & family history

Handwriting OCR is an AI service that turns handwritten documents into accurate, searchable, editable text. Unlike traditional OCR, which was built for printed text, it is built specifically for handwriting, including cursive, faded ink, historical scripts, and many languages, and it can translate non-English records into English in the same step. Upload a scan or photo and you get back clean text you can search, edit, and export to Word, Markdown, or plain text.

Handwriting OCR was founded in London in 2023, dedicated to applying modern AI to read the hardest handwritten documents: the cursive, faded, and historical pages that traditional OCR cannot handle. We are a small, independent UK team, the people who build the product also handle support, and we never use your documents to train models or share them with anyone.

Handwriting OCR is designed to process historical cursive styles including the flowing penmanship common in 19th-century documents. It handles connected letters, historical letter formations, and the stylistic variations typical of different time periods. Accuracy depends on the clarity of the original handwriting and scan quality. Well-preserved documents with relatively clear handwriting typically process well, while heavily faded or damaged sections may require more careful verification. The best way to assess performance on documents from your specific time period and region is to test with sample pages from your research.

Yes. Handwriting OCR processes scanned images and PDFs regardless of their source. If you can download or screenshot a genealogical record from an online archive, you can process it. The system handles various scan qualities and formats, including images downloaded from genealogy websites, photographs from archive visits, or personal scans of family documents. No format conversion or special preparation is required before processing.

Handwriting OCR extracts what is actually written in each document. If a surname appears as "Schmidt" in one census and "Schmitt" in another, the system will extract each spelling as written. This is actually valuable for genealogical research because it preserves the historical spelling variations that researchers need to track. However, you'll need to apply your own research judgment about whether variant spellings refer to the same family. The technology handles text extraction; genealogical interpretation remains researcher work.

Yes. Many genealogists use handwriting OCR specifically for this purpose. By processing family letters, diaries, census pages, and other handwritten documents, you create searchable text that you can organize into a personal research database. You can then search across your entire collection for names, dates, places, or events rather than manually reviewing each document. The extracted text can be exported in formats that work with genealogy software, personal databases, or research note systems.

No. Your documents remain private and are processed only to deliver results to you. They are not used to train AI models, not shared with third parties, and not retained longer than necessary to complete processing. This is particularly important for family documents that may contain personal information. Privacy is built into the service design as a fundamental principle, not an optional feature.

Yes. Handwriting OCR is designed to handle historical German scripts including Sütterlin, Kurrent, and earlier Gothic hands. These scripts present unique challenges with their distinctive letter formations and connected writing styles, but the technology is specifically built to process them. Accuracy depends on the handwriting quality and document condition. German family researchers working with 19th and early 20th century documents can process these materials, though verification against the originals remains important for genealogical accuracy.

Yes. Handwriting OCR processes historical French handwriting across different time periods, including parish registers, notarial records, and personal correspondence. It handles the cursive styles and abbreviations common in French genealogical documents. French-Canadian researchers, those working with Louisiana records, and anyone tracing French ancestry can process these materials. As with all historical documents, the quality of the original handwriting and scan affects accuracy, but the technology is built to work with the kinds of French documents that appear in genealogical research.

Yes. Handwriting OCR processes historical census records across different time periods, handling the cursive styles and formatting conventions used in different census years. Each decade used different schedules and formats, and enumerators wrote in the penmanship typical of their era. The technology adapts to these variations. Accuracy depends on the individual enumerator's handwriting and the condition of the census image, so the best way to assess performance for the years you research is to test with sample pages from those censuses.

Yes. Handwriting OCR processes military records from different conflicts and time periods, handling the cursive styles and documentation formats of each era. Revolutionary War records, Civil War compiled service files, and World War I draft registrations all used different administrative systems and conventions, and the technology adapts to them. It extracts military abbreviations and unit designations as written (for example "Co. K, 15th Reg't N.Y. Vol. Inf."), but interpreting unit organization, ranks, and reorganizations remains researcher work that requires military historical knowledge.

Yes. Handwriting OCR processes Latin text in parish registers, extracting names, dates, and ecclesiastical phrases. Many early registers used standardized Latin formulas for baptisms, marriages, and burials, and those repeated phrases help the technology recognize them accurately. You'll still need familiarity with ecclesiastical Latin to interpret the extracted text: understanding phrases like "baptizatus est" (was baptized) or "in matrimonium conjuncti sunt" (were joined in marriage) requires knowledge of church-record conventions, and dates expressed as feast days need conversion to the modern calendar.

Handwriting OCR extracts place names exactly as they appear in the manifest. If a birthplace is recorded as "Galicia," "Galizien," "Austrian Poland," or a phonetically spelled village name, the system extracts that exact text rather than standardizing it. This preserves the historical geographic terminology and spelling variants used at the time, which is valuable for immigration research. You then apply your own knowledge of historical geography and shifting political boundaries to interpret what these names meant and where they correspond on modern maps.

Handwriting OCR can process aged and faded documents, though accuracy depends on how legible the original remains. If you can read the text in the original photo or scan, the system will likely extract usable text. Severely faded sections may produce partial results that require verification against the original. Processing deteriorating baby books now creates value by capturing readable text before further degradation, so even where some sections need manual review you preserve content that might otherwise become completely illegible as the original continues to age.