GEDCOM MCP Server
Enables AI assistants to query genealogy data from GEDCOM files with 24 tools for search, navigation, relationship analysis, and timeline events.
README
GEDCOM MCP Server
A FastMCP server that enables AI assistants to query genealogy data from GEDCOM files.
Features
-
24 MCP Tools for comprehensive genealogy research:
Core Tools:
get_home_person- Get the tree owner's recordget_statistics- Tree stats (counts, date ranges, top surnames)get_individual- Get basic details by IDget_biography- Get comprehensive narrative package for one personget_family- Get family info (spouses, children, marriage)
Navigation Tools:
get_parents- Get parents of an individualget_children- Get all children from all marriagesget_spouses- Get all spouses with marriage detailsget_siblings- Get siblings (same parents)get_ancestors- Ancestor tree up to N generationsget_descendants- Descendant tree up to N generationstraverse- Generic graph traversal for custom navigation
Search & Discovery:
search_individuals- Search by name (partial match)semantic_search- Vector-based semantic search (e.g., "farmers in Scotland")search_nearby- GIS proximity search or bounding box search
Relationship Analysis:
get_relationship- Calculate relationships between two peopledetect_pedigree_collapse- Find ancestors appearing multiple timesfind_associates- FAN Club technique (Friends, Associates, Neighbors)
Timeline & Events:
get_timeline- Chronological life events for an individualget_military_service- Find all veterans in the tree
Place & Surname Analysis:
get_place_cluster- Get all people connected to a locationget_surname_origins- Analyze surname distribution and geographic origins
Natural Language:
query- Natural language questions (fallback for non-agent MCP clients)
-
4 MCP Resources:
gedcom://individual/{id}- Individual record by IDgedcom://family/{id}- Family record by IDgedcom://stats- Tree statisticsgedcom://surnames- All surnames with counts
-
Optimized for large files - Parses GEDCOM once at startup, builds in-memory indexes for fast search
Optional Features
Semantic Search Enable natural language semantic matching (e.g., "coal miners in Pennsylvania", "emigrated from Ireland"):
export SEMANTIC_SEARCH_ENABLED=true
Requires sentence-transformers (included in dependencies). First run builds embeddings (~15-30 seconds for large files), subsequent runs load from cache.
GIS Proximity Search Find people within X miles of a location or within a region's bounding box. Enabled by default with background geocoding. Results include:
- Proximity mode: Within X miles of a point
- Within mode: Inside a region's bounding box
No configuration needed - geocoding runs automatically at startup and caches results.
Telemetry & Observability OpenTelemetry tracing with Arize Phoenix for debugging and performance monitoring:
export PHOENIX_ENABLED=true
Requires Phoenix server running (see .env.example and docker-compose.yml).
Installation
Using uvx (recommended, no install needed)
uvx gedcom-server --gedcom-file /path/to/your/tree.ged
Using uv tool install
uv tool install gedcom-server
gedcom-server --gedcom-file /path/to/your/tree.ged
From source
git clone git@github.com:sjmatta/gedcom-mcp.git
cd gedcom-mcp
uv sync
Configuration
The server requires a GEDCOM file path. Optionally, you can specify a "home person" (the tree owner).
Configuration Options
| Method | Option | Description |
|---|---|---|
| CLI arg | --gedcom-file, -f |
Path to GEDCOM file |
| Env var | GEDCOM_FILE |
Path to GEDCOM file |
| CLI arg | --home-person, -p |
GEDCOM ID of home person (e.g., @I123@) |
| Env var | GEDCOM_HOME_PERSON_ID |
GEDCOM ID of home person |
CLI arguments override environment variables.
If --home-person is not specified, the server auto-detects the most connected individual in the tree.
Examples
# Using CLI arguments
gedcom-server --gedcom-file ~/Documents/family.ged
# Using environment variables
export GEDCOM_FILE=~/Documents/family.ged
gedcom-server
# With explicit home person
gedcom-server -f ~/Documents/family.ged -p @I500@
Claude Desktop Configuration
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):
Using uvx (recommended)
{
"mcpServers": {
"gedcom": {
"command": "uvx",
"args": ["gedcom-server"],
"env": {
"GEDCOM_FILE": "/path/to/your/tree.ged"
}
}
}
}
Using uv tool install
{
"mcpServers": {
"gedcom": {
"command": "gedcom-server",
"args": ["--gedcom-file", "/path/to/your/tree.ged"]
}
}
}
From source
{
"mcpServers": {
"gedcom": {
"command": "/path/to/gedcom-mcp/.venv/bin/python",
"args": ["-m", "gedcom_server"],
"env": {
"GEDCOM_FILE": "/path/to/your/tree.ged"
}
}
}
}
GEDCOM File
The server supports GEDCOM 5.5.1 format, which can be exported from:
- Ancestry.com
- FamilySearch
- MyHeritage
- Gramps
- Most other genealogy software
Development
# Install dev dependencies
uv sync
# Run tests
uv run poe test
# Run linter
uv run poe lint
# Run formatter
uv run poe format
# Run type checker
uv run poe typecheck
# Run all checks
uv run poe check
# Start server (requires GEDCOM_FILE env var)
uv run poe serve
For detailed development instructions, see CONTRIBUTING.md.
Pre-commit Hooks
Pre-commit hooks are configured for:
- Pre-commit: ruff linting, ruff formatting, mypy type checking
- Pre-push: pytest test suite
Install hooks:
uv run pre-commit install --install-hooks
License
MIT License - see LICENSE for details.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。