FamilySearch Genealogy Skill

Two modes of operation:

1. Live API (primary) — Query FamilySearch directly for person search, pedigree, descendants, historical records
2. Offline GEDCOM — Parse exported .ged files for local exploration, narratives, and statistics

Mode 1: FamilySearch API (Primary)

Prerequisites

1. FamilySearch Account — Free at \x3Chttps://www.familysearch.org>
2. Developer App Key — Register at \x3Chttps://www.familysearch.org/developers/>
3. OAuth 2.0 Access Token — Via authentication flow below

Store credentials:

Option A — Environment variable (all platforms):

export FAMILYSEARCH_TOKEN="\x3CTOKEN>"

Option B — macOS Keychain:

security add-generic-password -a "familysearch-app-key" -s "openclaw-familysearch" -w "\x3CAPP_KEY>"
security add-generic-password -a "familysearch-token" -s "openclaw-familysearch-token" -w "\x3CTOKEN>"

The script checks FAMILYSEARCH_TOKEN env var first, then falls back to macOS Keychain.

Authentication

OAuth 2.0 Authorization Code flow:

1. Direct user to: https://ident.familysearch.org/cis-web/oauth2/v3/authorization?response_type=code&client_id={APP_KEY}&redirect_uri={REDIRECT_URI}
2. User logs in, grants access
3. Exchange code for token:

curl -X POST "https://ident.familysearch.org/cis-web/oauth2/v3/token" \
  -d "grant_type=authorization_code&code={AUTH_CODE}&client_id={APP_KEY}"

4. Store returned access_token in Keychain

Sandbox for testing: https://integration.familysearch.org / https://api-integ.familysearch.org

API Usage

python scripts/familysearch.py \x3Ccommand> [args]

Search parameters: --given, --surname, --birth-date, --birth-place, --death-date, --death-place, --sex

Results use GEDCOM X format (JSON). Key fields: persons[].id, persons[].display.name, .birthDate, .birthPlace, .deathDate, .deathPlace.

Ahnentafel numbering in ancestry: 1=subject, 2=father, 3=mother, 4=paternal grandfather, etc.

Mode 2: Offline GEDCOM Files

For exported .ged files from FamilySearch, Ancestry, MyHeritage, etc. No API key needed — pure offline.

Getting a GEDCOM File

FamilySearch: familysearch.org → Family Tree → Tools → Export GEDCOM

Usage

python scripts/gedcom_query.py \x3Cgedcom_file> \x3Ccommand> [args...]

id_or_name: GEDCOM ID (e.g., I001) or partial name (fuzzy, case-insensitive).

Narrative Genealogy

Beyond data retrieval, this skill supports narrative genealogy — connecting facts to stories:

• Note occupations, migrations, life events when exploring a tree
• Cross-reference API/GEDCOM data with stories the user shares conversationally
• Build family narratives that explain why — not just dates and names
• Flag research opportunities: missing records, undocumented branches, conflicting dates
• Use story command (GEDCOM mode) for auto-generated biographical narratives

Agent Workflow

1. User asks about family history → Check if they have a FamilySearch account (API) or GEDCOM file (offline)
2. API mode: Search by name → get person ID → explore ancestry/descendants/relationships
3. GEDCOM mode: Load file → search → explore
4. Either mode: Combine structured data with user's oral history for richer narratives
5. Cross-reference: Use API to find records that fill gaps in GEDCOM data

Rate Limits & Best Practices

• FamilySearch API is free but rate-limited — cache results locally
• Never store FamilySearch usernames/passwords — OAuth tokens only
• Tokens expire; re-authenticate on 401 responses
• GEDCOM parser handles files up to ~100K individuals
• File encoding: auto-detects UTF-8 (with BOM), UTF-8, latin-1