Genealogy research toolkit

What Genealogy research toolkit MCP server does

Genealogy research toolkit MCP server supports research across 114 archives and databases in 55 countries. It helps identify which records may exist for a place and period, where historical records are held today, and how to search them. Coverage includes regions such as Europe, Latin America, the Russian Empire, the former USSR, the British Isles, and the United States.

The toolkit also handles common interpretation problems in historical genealogy. It can generate spelling variants, transliterations, old-orthography forms, gendered names, patronymics, given-name equivalents, and Daitch–Mokotoff codes. Date tools normalize uncertain or multilingual dates, convert between calendars, interpret feast-day references, and read Church Slavonic numerals.

How it works

The server runs locally as an MCP process and can be started with uvx genealogy-mcp. An MCP-capable assistant sends research questions to individual tools, which return plans, source information, calculations, links, or assessments. Online searches are performed only when requested and target the selected service.

Research can begin with a country, region, time range, record type, or known person. The assistant can then combine source discovery with name variants, historical place information, date conversion, and evidence comparison. check_match compares facts in a candidate record, while audit_tree checks a GEDCOM or family_tree.json for contradictions, duplicates, and missing information.

Tools and capabilities

The Genealogy research toolkit MCP server exposes tools for:

• Planning research and finding sources by geography, period, and record kind.
• Explaining source coverage, access methods, and search approaches.
• Identifying present-day archives for records from historical provinces.
• Creating search links and searching Internet Archive full text, Chronicling America, WikiTree, Open Archives, and Open List.
• Looking up historical place affiliations, coordinates, and related Wikipedia pages through Wikidata.
• Converting names, dates, calendars, feast dates, and Cyrillic numerals.
• Calculating kinship, expected DNA shares, and relationships compatible with shared centimorgans.
• Comparing a possible record with a person and auditing family-tree data.
• Drafting archive or registry requests and explaining record terminology.

Setup and configuration

Claude Desktop can use the downloadable .mcpb release package. Claude Code can register the server with claude mcp add genealogy -- uvx genealogy-mcp. Other listed MCP clients can use uvx genealogy-mcp as the server command, subject to their MCP configuration format.

The repository provides guides in English, Russian, and Ukrainian, with major guides also available in Spanish, Portuguese, Polish, and German. No environment variables are specified in the provided setup instructions.

Limitations and notes

The toolkit does not establish identity from a matching name alone. Its matching workflow weighs additional facts and reports a verdict, but documentary evidence remains necessary. DNA calculations describe relationships that fit a shared-cM value; they do not identify the relationship by themselves.

Online search depends on the selected service and is performed only when requested. Historical borders, spelling systems, calendars, and archive transfers can make results incomplete or ambiguous. Files that are not UTF-8 are read-only according to the project documentation.