Scans in shoeboxes, dates in a grandmother's notebook, a tree in MyHeritage, stories that only one aunt still remembers. gene-archive gathers all of it into one beautiful website your whole family can open on a phone: an interactive tree, a card for every person with the documents behind each date, an archive of scans with transcriptions, a map of family places and the story of the family — in English, Russian and Romanian.

Relatives don't just read it. They leave comments, suggest corrections and, with an account, edit the archive right in the browser. You keep everything: one JSON file, your own domain, free hosting, full history in git, GEDCOM in and out.

Grew out of a real family archive: ~550 people, ~500 documents, three countries, relatives commenting in three languages. The demo is a fictional family made to show every feature.

Why gene-archive

Features

🌳 Tree, people and "How we are related"

• Hourglass tree around any person, a whole-tree chart, export to PNG, branches by surname.
• Person card: dates and places with their documents, alternative versions of a date, relatives, a connected life story.
• How we are related: the exact relation between any two people — second cousin once removed, great-uncle, half-brother, in-laws — with the chain of people in between. In Russian and Romanian too, with their own kinship terms.
• Search everything from the magnifier: names, villages, any word of a document transcription, places, the story, leads — word endings, the pre-1918 spelling and Latin/Cyrillic letters do not get in the way (“Kolodin” finds Колодин).
• People: a calendar of birthdays and anniversaries, surname pages, the zodiac.

🗂 Archive of documents

• Every scan and photograph with a transcription, the back of a photo, PDF previews.
• Filters by type of document and by family line; restored photos next to originals.
• Sources: every archive and database you searched — what it is, what was found there, the documents and the research-log entries for each; plus a log of corrections.

🗺 Places and the family story

• A map (Leaflet + OpenStreetMap) with family branches, moves between places, "then and now" names and timelines.
• The Story page: your narrative with a timeline and a "by place" filter.
• Questions to relatives: what the family does not know yet — answered right on the site.

✏️ Online editor

• People, families, documents (with scan uploads) and places in a browser — on a phone too.
• Dates the way people write them: "about 1899", "between 1900 and 1905", "9.10.1899".
• A change that would break the data is not saved; the editor says what is wrong.
• Two people editing the same record? The second one is warned instead of overwriting.
• Every change goes to the corrections log and becomes a git commit by its author; an administrator can restore any version.
• Suggest a correction: readers send a fix from a person's card; an administrator applies it in one click.

🔎 Research assistant

• Data checks (gene hints and the editor's "Checks" tab): impossible facts first — born after the mother's death, a parent aged 11 — then gaps: no dates, a date without a document, people not connected to the tree, documents nobody refers to.
• A step-by-step research guide and ready AI prompts for archives and open sources.

🚀 Publish anywhere

• gene build → plain HTML for GitHub Pages, Netlify, Cloudflare Pages or any web server.
• gene serve or the Docker image adds comments, a PIN for a private archive and the editor.
• Share any person, document or place: the link shows a photo, a name and a description in Telegram, WhatsApp and Facebook.
• Search engines find every person and surname: static pages, sitemap, Schema.org; living people are noindex.

How it compares

gene-archive does not try to be a database of the world's records. It is the place where your family's evidence and stories live — and where your relatives actually look.

Quick start

You need Python 3.10+. Pillow makes thumbnails (optional), pdftoppm from poppler-utils makes PDF previews (optional).

pip install "gene-archive[images] @ git+https://github.com/igormel81/gene_archive"

gene init my-family --example     # a copy of the demo family to look around
gene serve my-family              # → http://127.0.0.1:8111

Start your own archive:

gene init my-family --lang en     # an empty site: family_tree.json + site.json
gene import tree.ged my-family    # …or start from a GEDCOM file
gene validate my-family           # check the data
gene build my-family              # the site is in my-family/_site

Then fill in family_tree.json (people, families, documents, places — see the data format) and site.json (settings), put scans into sources/, and write the story in content/story.html (content pages).

New to genealogy? Read How to start researching your family history: interviews, home documents, online databases, archives, evidence and how to record it.

Quick start with an AI assistant

Works in any AI coding environment — Claude Code, Codex, Cursor, GitHub Copilot, Windsurf, Gemini CLI. Open an empty folder in it and paste:

Set up a family archive site with gene-archive (https://github.com/igormel81/gene_archive) in this folder.
1. Install it: pip install "gene-archive[images] @ git+https://github.com/igormel81/gene_archive"
2. Create the site: gene init . --lang en (or gene import <my file>.ged . if I give you a GEDCOM file),
   then git init and a first commit.
3. Read AGENTS.md and follow it in all further work: never invent facts or relationships,
   cite a document for every fact, keep doubtful things as leads.
4. Ask me about the family (my parents, grandparents, what documents and photos we have at home)
   and fill family_tree.json and site.json from my answers.
5. Run gene validate and gene serve ., give me the link and show what you changed.

Then just talk to it: "here is a scan of my grandfather's birth certificate", "my aunt says…". More ready-made requests and the rules for reviewing the agent's work: docs/ai-assistants.md.

Documentation

Commands

Messages are in Russian when the system language is Russian (or GENE_LANG=ru).

License

MIT. Leaflet is BSD-licensed, map tiles © OpenStreetMap contributors — see THIRD_PARTY.md.