io.github.igormel81由 GitHub 帳號 igormel81 發布v0.7.0更新於 Oct 30, 2026
Family history: 171 sources, book search, calendars, kinship, event evidence and research projects
家譜研究工具包，讓助理取得 171 個檔案來源、姓名與曆法工具、證據查核，以及可長期保存的研究專案。
Desktop only，透過 STDIO。 STDIO 服務會啟動本機處理程序，因此需要 SourceWeft 桌面主機。
參照 儲存庫 中的啟動說明。
[CI] [PyPI] [Python] [MCP Registry] [License: MIT]
An MCP server that turns Claude, ChatGPT, Cursor or any MCP-capable assistant into a careful genealogist. It knows where the records of your ancestors are (171 archives and databases across Russia and the former USSR, Europe, the British Isles, the USA and Latin America), how to search them (name variants, transliterations, old spelling, the Julian calendar), and when a find really is your person — so the assistant gives you working links and a plan instead of confident guesses.
Built from years of real research for an open family archive powered by gene_archive: parish registers, revision lists, WWI and WWII databases, repression files, emigration records — and every trap we fell into along the way. The story behind it: an article on Habr (in Russian).
.ged file from MyHeritage, Ancestry, Gramps or RootsMagic
and it reads, searches and changes it in place — only the changed lines, everything else byte for byte, a copy kept,
every change shown first. Or at a gene-archive family website: changes go
through the site's checks, logs and git history.Claude Desktop: download genealogy-mcp-<version>.mcpb from the
latest release and open it.
Claude Code: claude mcp add genealogy -- uvx genealogy-mcp.
Anything else: uvx genealogy-mcp as the server command (details in Install).
Then just ask: "Help me find my great-grandfather. Here is what I know…"
My great-grandfather Ivan Kolodin was born about 1885 near Bogoroditsk, Tula province. Where do I look?
The assistant makes a plan for Russia 1880–1945, reads the record guides, builds the name forms (Колодин / Колодина / Колодинъ / Kolodin / Kołodin…), gives you search links for FamilySearch, the Internet Archive full text and the Russian war databases, tells you which archive holds the parish registers and drafts the request.
Is this baptism record the same Ivan? It says «Іоаннъ, father Петръ, 1887, Сынково».
check_match compares fact by fact with name variants and age tolerance: two independent facts beyond the name →
confirmed; one → probable; a contradiction → find a third document.
What is my grandmother's cousin's daughter to me? И как это по-русски?
kinship → «троюродная сестра… / first cousin once removed» — in Russian, English and Romanian, with the expected DNA share.
Check my tree. (a GEDCOM from Ancestry, MyHeritage, Gramps, Geni…)
audit_tree finds the impossible (born after the mother's death, a mother aged 10, death before birth), possible
duplicates and gaps — a to-do list for the next search.
The register says he was born «на Покров» 1885, the year is written ҂аѡ҃пе.
cyrillic_numerals reads the year (1885), feast_date gives 1 October (Julian) = 13 October 1885.
We share 230 cM with a match on MyHeritage. Who could she be?
dna_match: second cousin, first cousin twice removed, half first cousin… — and why only documents can decide.
Where are the parish registers of Volhynia governorate now?
where_are_records: the state archives of Zhytomyr (the consistory), Rivne and Volyn — borders moved, so it is not
always the archive of today's region.
Remember what we found, and continue next week.
A research project on your computer: people, sources with full citations, open leads and a log of every search — including "nothing found" — so nobody searches the same place twice. GEDCOM in and out.
Write what we found into our family site.
If your family archive is a gene-archive site, give the project tools the path of its folder: the assistant reads the archive itself — who is missing, which dates have no document — and every change it makes goes through the site's own checks, lands in its corrections and research logs, is committed to git under your name and rebuilds the site. It shows the change first and writes only after your yes.
Here is my tree — add what we found.
Or give it the GEDCOM file of the tree you keep in MyHeritage, Ancestry, Gramps, RootsMagic, Family Historian or webtrees: the assistant reads it, searches it and changes it in place. Only the lines of what changed are rewritten — your program's own tags, photos, notes and the header stay byte for byte; it shows those lines before writing, keeps a copy of the file, and keeps the search log, leads and corrections in a file next to it. Then import the file back into your program (Family Historian and webtrees work on GEDCOM directly). Files not in UTF-8 are read-only.
Most genealogy MCP servers connect the assistant to one service or one file. genealogy-mcp gives it the research method around them — where to look, how to search, how to read and weigh what it finds — and works on your own tree file too. It runs happily next to a server for a service you use, such as a FamilySearch client.
Everything runs on your computer. The server goes online only when the assistant runs a live search
(search_online, search_book_pages, place_history), and then sends that query to that one site.
Record guides in English, Russian and Ukrainian, the main ones also in Spanish, Portuguese, Polish and German.
Claude Desktop, one click: download genealogy-mcp-<version>.mcpb from the
latest release and open it — Claude Desktop installs
it and asks where to keep your research projects.
For the other ways you need uv (it fetches Python and the package by itself).
Claude Desktop, by hand — Settings → Developer → Edit Config, add:
Claude Code:
Cursor (~/.cursor/mcp.json), Windsurf, Cline — the same mcpServers block as for Claude Desktop.
VS Code (.vscode/mcp.json): {"servers": {"genealogy": {"command": "uvx", "args": [ …same… ]}}}.
Docker: docker build -t genealogy-mcp . then docker run -i --rm genealogy-mcp (stdio) or
docker run --rm -p 8000:8000 genealogy-mcp --http --host 0.0.0.0 (HTTP at /mcp).
Your gene-archive site as the project: install the server together with gene-archive —
pip install "genealogy-mcp[site]" or uvx --from "genealogy-mcp[site]" genealogy-mcp (the gene command is found
on PATH or next to the server's Python; elsewhere set GENE_CMD), then name the folder in GENEALOGY_MCP_SITES (or in the Claude Desktop extension settings), or just give its path
to the assistant. GENE_AUTHOR sets the name in the git history (default: git config user.name).
Your tree in a GEDCOM file as the project: give the assistant the path of the .ged file (or name it in
GENEALOGY_MCP_SITES). Copies are kept as tree.ged.bak (before the last change) and tree.before-genealogy-mcp.ged
(before the first); the log lives in tree.ged.research.json.
Clients that connect over HTTP (ChatGPT connectors, remote agents): run
genealogy-mcp --http --host 0.0.0.0 --port 8000 behind HTTPS and add https://your-host/mcp.
In HTTP mode audit_tree takes the GEDCOM text only, never a file path.
The package is on PyPI and in the official MCP Registry (io.github.igormel81/genealogy-mcp): uvx genealogy-mcp installs and runs it.
Start from project_view: policy, local sources, active questions and owner tasks. Work
from family documents and the civil registry through existing records to primary sources,
then secondary leads. project_policy configures limits and archive-request preferences
per project. Leads have a next step; source checks distinguish no hits from an inaccessible
site. Custom archive folders without site.json are readable without converting or
changing them. See workflow, policy and source checks.
The 171 catalogue entries are recommendations, not 171 live connectors. Six services
support search_online; Wikidata supplies place history. Other sources need a browser
or documents supplied by the user. Gallica page search and event evidence are described in
books and event citations.
Resources: genealogy://sources (the catalogue), genealogy://guide/research/{en|ru} (a full research guide),
genealogy://guide/prompts/{en|ru} (ready requests for each step), genealogy://records/{topic}, genealogy://glossary.
Prompts: start research, check a find, transcribe a record, write to an archive.
search_online, search_book_pages, place_history) sends the query — names,
years, place — to the one site searched. Research projects are files on your computer
(~/Documents/genealogy-mcp or GENEALOGY_MCP_HOME); on a shared HTTP server they are switched off.evals/tasks.json holds 35 questions with checkable answers: dates by church feasts, Julian, Hebrew and Republican
calendars, Cyrillic numerals, Daitch–Mokotoff, Moldovan surnames, DNA, kinship, old terms, archives of historical
provinces and borders by year, evidence checks, ages that cannot be one person,
record history. evals/check_tools.py runs in CI and checks that the tools answer all of them correctly.
evals/run.py asks a Claude model the same questions without the server and with it and scores both
(needs an API key; it costs money, so it is run by hand: Actions → Benchmark).
See CONTRIBUTING.md for source fields and contribution guidelines.
A source is missing or a link moved? Edit tools/make_sources.py, run python3 tools/make_sources.py and
python3 -m unittest discover tests, open a pull request. Record guides live in src/genealogy_mcp/guides.py.
MIT License.
來源：README.md，提交 7f4e96b
Find genealogy databases and archives for a country (or region), period and kind of records — 171 sources across Russia and the former USSR, the Caucasus and Central Asia, Eastern, Western and Northern Europe, Germany, the British Isles, the Americas, Australia and New Zealand, each with what it holds and how to search it well.
langtextlimitcountryrecordsyear_toyear_frominclude_worldwideEverything about one source from the catalogue (by id from find_sources): coverage, access, search tips.
langsource_id必填Ready search links (open them in a browser) in the sources that take a search URL — FamilySearch, Internet Archive full text, Find a Grave, Arolsen, Geneteka, Open List and others — for a name, with the name's variants.
givenplacecountrysurnameyear_tovariantsyear_fromWhich archives now hold the records (parish registers, revision lists, confession lists, civil records) of a province of the Russian Empire, the Kingdom of Poland, Galicia, Bukovina, East Prussia, the Baltics or the Caucasus — borders moved, so this is often not the archive of today's region. Also works from a modern region back to the historical provinces it was part of. With a year, says whose land it was then (southern Bessarabia — Cahul, Ismail, Bolgrad — was Moldavian/Romanian in 1857–1878, all Bessarabia Romanian in 1918–1940 and 1941–1944); with a village, which church kept its books (a village without a church was recorded in a neighbouring one).
langyearplace必填villageSearch an open database live and get the matching records back (title, date, place, link, snippet). Sends the query to that one site only. Run it for each name form from name_variants; for FamilySearch, Geneteka, Find a Grave and other sites without an open API use search_links.
givenlimitplacequerystartsource必填surnameyear_toyear_fromFind OCR excerpts inside one Gallica book using its ARK/document_id from search_online source='gallica'. Returns scan sequence numbers and stable page URLs, not verified facts or printed page numbers. Sends the query only to Gallica's documented ContentSearch API. start is a zero-based result offset. Access errors are incomplete checks, never no hits; no OCR hit does not prove absence from the scans.
limitquery必填startdocument_id必填Look a place up in Wikidata: what administrative units it was part of (with years where known), coordinates and the Wikipedia article — a first step to the uezd, province and parish of a village — and the church its books were kept in, where that attachment is known. Sends the name to Wikidata.
langname必填A step-by-step research plan for a country and period: which records exist, which guides to read and which sources to search, in order (free and country-specific first).
langcountrylead_idprojectyear_toquestionyear_fromA plan for reading unindexed FamilySearch films page by page for a parish, years and parts: which films to read in which order, thumbnails to find the book (title pages, end-of-year summaries naming the church) and full frames only for the record pages, a rough request budget split into sessions under the site's variable rate limit (pause at least an hour after HTTP 429, no parallel sessions), and what to log with project_coverage.
langfilmsorderpartsparish必填projectyear_toyear_from必填All the forms a name can take in records (and its Daitch–Mokotoff Soundex codes): spelling variants, pre-1918 orthography, male/female surname forms (Колодин/Колодина, Kowalski/Kowalska, Kazlauskas/Kazlauskienė), transliterations (passport, BGN, German, French, Polish, Romanian, Spanish, Ukrainian, scholarly) and given-name equivalents across languages (Иван = Іоаннъ = Jan = Johann = Juan). Moldovan and Bessarabian surnames: Romanian and Russian renderings (Călugăreanu = Калугарян = Калугаряну = Кэлугэряну, Pleș = Плешъ = Плешу), a second surname flagged («Мокану, он же Плеш»), clerk and recognition distortions as hints, and look-alikes that are another family.
kindname必填originA date as people write it ('около 1899', 'between 1900 and 1905', '9 de octubre de 1899', '9.10.1899') → ISO or GEDCOM ('ABT 1899', 'BET 1900 AND 1905', '9 OCT 1899', '1899-10-09').
text必填Convert between calendars of old records: Julian (old style; Russia until February 1918: +10 days in the 1600s, +11 in the 1700s, +12 in the 1800s, +13 in 1900–1918), Gregorian, Hebrew (Jewish registers, tombstones) and the French Republican calendar (1792–1805).
todate必填from_calendarThe date of a church feast or fast in a given year (Julian and Gregorian), including movable feasts counted from Easter. Records and family stories often date events by feasts: «родился на Покров», «венчались после Петрова поста».
year必填feast必填churchChurch Slavonic numerals in old registers and church books: ҂аѡ҃п = 1880, ѳі = 19 (11–19 are written units first). Reads letters, writes a number, or finds every numeral marked with a titlo in a text.
value必填Which relationships a DNA match fits, from the amount of shared DNA (Shared cM Project 4.0 ranges), closest first. Several always fit: use documents (and the kinship tool) to decide.
langshared_cm必填The birth-year range implied by an age in a record (census, confession list, death record), exact and loose.
age必填record_date必填A birth year estimated from relatives (eldest child −27, spouse and siblings ±8, parents +18…+45, first marriage at about 22) and ages in records: a GEDCOM date (BET … AND … or ABT …), the most likely year and a ready explanation to save with it. Also checks the ages against each other: 7 in 1850 and 22 in 1882 cannot be one person — likely two namesakes.
sexageslangfather_birthmother_birthspouse_birthfirst_marriagesiblings_birthseldest_child_birthWhat B is to A, named in Russian, English and Romanian (троюродный брат / second cousin / văr de gradul 2, двоюродный дед / great-uncle), with the expected share of DNA.
upsexdownhalfpathIs the person in a found record ours? Compares fact by fact (with name variants and Julian/rounded-age tolerance) and gives a verdict: confirmed (2+ independent facts beyond the name), probable, insufficient, conflict or different_person.
langfound必填known必填year_toleranceA working hypothesis about a parent from a patronymic: the father's given name it implies (Петрович → Пётр, Ильич → Илья), each candidate ranked by name, surname, village and age with the arguments for and against (a grandson named after him, godparents and sureties of his surname, his household listing a child of this name), a warning for a namesake from another village, and the record that would settle it. Always a hypothesis — save it as an open lead or 'probably', never as a link in the tree.
langchild必填candidatesAudit a family tree from any program (GEDCOM) for impossible facts (born after the mother's death, a parent under 14, death before birth), possible duplicates and gaps (no sources, no dates, people cut off from the tree) — a to-do list for the next search.
file必填langcodeslimitA draft letter to a registry office or an archive (Russian ZAGS, state, military and FSB archives; any archive in English, Spanish or Portuguese), with what to attach. Gaps are left in [brackets]. The user sends it — not the tool.
borndiedkind必填whatplaceyearspersonspousearchivedetailsparentsproject另有 5 個How to read and find a kind of record: parish registers, confession lists, revision lists, household books, WWI and WWII records, repression files, passenger lists, censuses, European civil registration, Latin America, evaluating evidence — in English, Russian, Ukrainian, and partly Spanish, Portuguese, Polish, German. Without a topic — the list.
langtopicOld Russian terms from records — estates, kinship by marriage (шурин, деверь, золовка), archive citation (фонд, опись, дело, лист), administrative units, pre-1918 letters — explained.
langtermRead the research project kept between conversations — or a family's gene-archive site folder, or the user's own tree in a GEDCOM file: summary (people, open leads, recent searches, problems), one record with its family and sources, a text search, the full search log, or the tree checks. Start every session on an existing project with what='summary'.
idlangwhatqueryprojectAdd or change a person, family, source or research lead in the project. Every change of an existing record is logged with the old and new value, the reason and the documents. Save the source first, then cite its S-id. In a gene-archive site folder the change goes through `gene apply`: refused if it adds an error, logged, committed to git under the owner's name, and the site is rebuilt. In a .ged file only the changed lines are rewritten (everything else in the file stays as it was), a copy is kept, and gedcom_changes shows the lines. Ask the user before writing into their archive or tree.
kind必填deletereasonrecord必填dry_runproject必填source_idsReview source references for individual birth/death/burial/residence and marriage events, read-only. Uses the project root (or root_person_id), follows ancestors up to 10 generations and prioritizes closer ones. Distinguishes missing/broken references, missing locators and reference presence. A person's general sources do not prove every event; even a located citation still needs inspection. Does not verify parentage or change data.
limitproject必填generationsroot_person_idConfigure this project's policy only (dry_run by default; save within the owner's authorization). Keys: max_active_questions (default 3), allow_archive_requests (default true; drafts only), allow_living_research (default false), owner_tasks [{id, next, due (ISO date), status}]. Read policy, resources and waiting questions with project_view what='workflow'. A custom archive folder without site.json stays read-only; its policy is stored under GENEALOGY_MCP_HOME/policies, never in that folder.
changes必填dry_runproject必填Write a line in the project's research log, so nobody searches the same place twice. Log negative results. Use source_id/lead_id to track the source checklist; unavailable is not no hits.
scopeaction必填result必填lead_idoutcomeproject必填source_idWhat has already been read in the registers: parish → year → part → film and frames → status (complete, partial up to No. …, not found). Mark each book when you finish it, and ask what is still unread for these villages and years before opening a film — so nobody reads the same frames twice. Kept with the project (for a .ged file in the file next to it; for a gene-archive site in a file of this server, read together with the site's own coverage).
langpartsactionplacesdry_runentriesproject必填year_toyear_fromExport the project as GEDCOM 5.5.1 (for MyHeritage, Ancestry, Gramps, Geni…) next to the project file, or start a new project from a GEDCOM exported by another program.
fileproject必填direction必填已滿足適用檢查項的 90%
該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。
請在註冊中心條目中宣告憑證請求標頭，或保持端點能夠回應探測，以便判定其認證方式。
{ "mcpServers": { "genealogy": { "command": "uvx", "args": ["genealogy-mcp"] } }}claude mcp add genealogy -- uvx genealogy-mcp