Immigration documents in genealogy: how to extract names, dates and places

Several types of documents were generated during the immigration process, each serving a different administrative purpose but all potentially valuable for genealogical research:

• Passenger lists (ship manifests) — compiled at the port of departure or arrival, listing every person aboard. These are the most common immigration records and often the most detailed.
• Colony entrance records — in countries like Brazil and Argentina, immigration colonies maintained their own registers of arriving families, including lot assignments and family composition.
• Work contracts — in some immigration programs (especially Brazilian coffee plantation immigration), contracts recorded the worker's name, origin, family members and destination farm.
• Passports — internal and international passports from the 19th century contain identity information, physical descriptions and sometimes the specific reason for travel.
• Naturalization records — when an immigrant became a citizen of the new country, the process generated documents recording their original nationality, date of arrival and sometimes family information.

Immigration documents, particularly passenger lists, contain a rich set of fields that are directly useful for genealogical research:

The town of origin is the single most valuable field in any immigration document. While a name and a date narrow the search, it is the specific town — Glandorf, not just Hannover; Molinella, not just Bologna — that makes it possible to locate the original birth or baptism record in a European archive.

With the town of origin identified, the researcher can search parish records, civil registries and state archives in the specific municipality where the ancestor was born. This is the bridge between the immigrant family in the Americas and the original family in Europe — often the breakthrough moment in a genealogical investigation.

Not all immigration records include the town of origin. Some list only the country or region (e.g. "Prussia" or "Kingdom of Sardinia"). In these cases, complementary sources — marriage records in the destination country, parish records mentioning the parents' origin, or other immigrants from the same ship — may help narrow down the location.

Immigration documents are particularly well-suited for AI-based extraction because of their structured, tabular format:

• Tabular structure — passenger lists are organized in columns with consistent headers, making field identification more reliable than in free-form parish records.
• Surname search with variant tolerance — AI extraction can match target surnames even when misspelled by port agents, using fuzzy matching with Levenshtein distance and known variant mappings.
• Bulk PDF processing — entire immigration books with hundreds of pages can be processed in a single upload, with the system extracting only the records matching the target surnames.

KinProof processes immigration documents by combining OCR (for printed or typewritten lists) with AI vision models (for handwritten manifests), then applying surname matching across every page to find all occurrences of the target family.