Genealogy GEDCOM & File Formats

Every genealogy app reads and writes GEDCOM. None of them does it perfectly. Understanding the format prevents data loss when you move between Genealogy Online Services, Genealogy Desktop Software, and Genealogy Self-Host & Web Software. For DNA file formats see Genealogy DNA Third-Party Analysis.

The GEDCOM standards

• ★ ★ GEDCOM 5.5.1 — the de-facto standard since 1999. Plain-text, line-numbered records (0 @I1@ INDI, 1 NAME John /Smith/), CONT/CONC line-folding, ASCII / ANSEL / UTF-8 encodings (cursed: ANSEL is a custom mapping no modern editor handles cleanly). Universal but quirky — every app implements a slightly-different dialect, and round-tripping loses something. The safe lowest-common-denominator export in 2026.
• ★ GEDCOM 7.0 — the 2021 modernization by FamilySearch. UTF-8 only, formal data model, structured sources, cleaner extension mechanism, supports _LANG / _GENDER_IDENTITY / non-binary, multi-page citations. Adoption growing 2024–2026: Gramps, RootsMagic, Family Historian, webtrees, MacFamilyTree all import; export support varies. Use 7.0 between modern apps; fall back to 5.5.1 for ancient ones.
  • https://gedcom.io/
• GEDCOM 5.5 (1996) — predecessor; only encounter when reading very old files.
• GEDCOM 4.x and earlier — historical; PAF-era. Convert via Gramps or RootsMagic if you find one.

Adjacent / specialized formats

• GEDCOM-X — JSON / XML / Atom-based modern model from FamilySearch (2012-ish). Used by FamilySearch's own APIs, not by third-party apps. Don't confuse with GEDCOM 7.
• FamilySearch FS-API — REST API; paid commercial partners + free for non-commercial open-source apps with approval. RootsMagic, Family Tree Maker, Ancestral Quest, Heredis sync via this.
• GEDML — XML-flavored GEDCOM variant; ~zero adoption.
• GenealogyXML / GeneanetXML — vendor-specific; export options in some apps.
• GRAMPS XML — Gramps' native portable format; round-trips losslessly with itself, partially with others.
• Heredis HMW, MacFamilyTree FTM2, Family Tree Maker .ftm, Reunion Family File — each app's binary native format. Use these only on the originating app; export GEDCOM for archive / portability.
• TMG SQZ — The Master Genealogist's bundle (sunset 2014); migrate via Second Site or Family Historian's TMG-to-GEDCOM importer.

Conversion / inspection / cleanup tools

• ★ GedSite GEDCOM Inspector — free GEDCOM analyzer; surfaces validation errors, statistics, encoding issues.
• ★ gedcompare — free; diff two GEDCOM files; useful when re-importing after edits.
• Gramps — the de-facto best GEDCOM parser among FOSS apps; can read messy files other apps reject and repair them.
• GED-inline — visual line-level inspector.
• GEDCOM Validator (gedcom7.io) — free 7.0 validator, also handles 5.5.1.
• RootsMagic GEDCOM Mover — free Windows utility for messy file repair.
• GEDmill — free open-source GEDCOM-to-HTML site generator.
• Visual Chartform — paid; visualize relationships from a GEDCOM file.
• Geneanet Open — free uploader; one-way export from desktop apps to Geneanet web tree.

Common GEDCOM round-trip pitfalls

• Custom event types — GEDCOM 5.5.1 has a fixed-ish vocabulary; custom events become _TYPE extensions that vary by app and may be silently dropped on import. GEDCOM 7's structured extensions help.
• Source citation depth — Evidence Explained-style sourcing in RootsMagic / Gramps loses fidelity in any GEDCOM round-trip; the app re-imports as flatter sources.
• Media file paths — relative OBJE FILE paths break on import to a different folder structure. Pack media as a tarball alongside.
• Non-Latin scripts — UTF-8 GEDCOM is fine; ANSEL-encoded files often corrupt. Convert to UTF-8 in Gramps before further moves.
• Living-person privacy — most apps offer "GEDCOM export with living people privatized." Use it when sharing publicly.
• Notes vs. research notes vs. "research log" — non-standard; varies by app. Plain NOTE records are the safest target.

What's changing in 2024–2026

• GEDCOM 7 momentum — FamilySearch published reference parsers; webtrees, Gramps, RootsMagic shipped readers; Family Tree Maker partial. Round-tripping 7 → 5.5.1 still loses metadata.
• JSON-y sidecar formats — some apps adopting JSON-LD for citations; not standardized.
• AI-assisted GEDCOM cleanup — community tools using LLMs to repair encoding, normalize place names, deduplicate individuals on import.

Pick this format if…

• Sharing with the most apps: GEDCOM 5.5.1 UTF-8.
• Modern app to modern app: GEDCOM 7.0.
• Lossless backup of your own tree: the app's native format plus GEDCOM 5.5.1 export plus a media tarball.
• Publishing as a static site: GedSite or GedHTree from a 5.5.1 file.
• Validating a messy file: Gramps (most forgiving parser) → re-export clean GEDCOM.