AI Genealogy Tool Tackles Soviet-Era Family Histories with Evidence‑First Approach

Overview

The open‑source skill, hosted at https://github.com/DenisSergeevitch/ex-ussr-kinship-research turns an AI agent into a meticulous genealogist for the former USSR, Soviet Union and Russian Empire. It reads church registers, civil acts, military files, deportation and repression documents, and unindexed archive scans. Every discovery is linked to a source file, and the system distinguishes between proven facts, reported claims and unresolved items.

Core Capabilities

• Evidence‑first protocol – each claim receives a status such as hypothesis, single‑source, corroborated, strongly‑corroborated, contradicted or unresolved. Mirrors of the same record share an independence group, preventing double counting.
• Historical jurisdiction model – events are mapped to the administrative unit that existed at the time of the event, not to modern borders.
• Name‑variant matrices – the tool recognises pre‑reform Russian, Ukrainian, Polish, Romanian, German and Yiddish spellings, patronymics, maiden names and transliterations.
• Reproducible registers – persons, artifacts, claims, relationships and a full CSV search log are stored; a Markdown report template can be generated.
• Archive request workflow – the skill can draft and, when authorized, send focused inquiries to archives, libraries or museums, recording replies as evidence.

Practical Implications

Researchers can pose a specific question, for example “My great‑grandfather was exiled to Kazakhstan in the 1930s; locate his parents in church records.” The agent searches relevant FamilySearch books, extracts permitted fragments, and builds a chain of evidence. The final HTML report shows each relationship edge, any contradictions, failed searches and concrete next steps. PDF export is optional and requires a local browser.

Verification Process

To confirm the announcement, check the repository’s README, the list of supported republics (all 15 former Soviet states), and the evidence‑protocol description. The skill’s audit script validates referential integrity and flags weak corroboration before conclusions are drawn. Users should review the generated search log and ensure that any claimed independence group corresponds to a cited artifact.

Limitations

The skill does not retrieve personal data of living individuals, nor does it bypass privacy protections such as CAPTCHAs or paywalls. It requires the user to supply authorized access to archives or to approve institutional correspondence. Success depends on the availability of original records; many regions have sparse documentation, which may leave some queries unresolved. The audit script’s reliability is not publicly documented, so its referential integrity checks remain unverified. GEDCOM file creation is user‑initiated, and the tool does not automatically synchronize external trees.