This project provides a structured framework of prompts, vault templates, and archive guides for AI-assisted genealogy research, designed to accelerate family history work while maintaining source rigor. It targets genealogy researchers seeking to leverage AI and AI/tech enthusiasts interested in autonomous research applications, offering a concrete example of generalized methods derived from extensive personal research.

How It Works

The core approach employs structured, autonomous research loops, inspired by Andrej Karpathy's concept, adapted for the nuanced domain of genealogy. AI prompts are meticulously defined with goals, metrics, verification conditions, and guardrails. The system autonomously searches the web, updates a markdown-based vault, and critically, performs mechanical verification through cross-referencing independent sources. This methodology emphasizes logging negative results and maintaining probabilistic confidence tiers (e.g., Strong Signal, Moderate Signal, Speculative) rather than treating all data as equal, acknowledging the often conflicting nature of historical records.

Quick Start & Requirements

Clone the repository. Copy the vault-template/ folder into your preferred markdown editor (e.g., Obsidian). Populate Family_Tree.md with existing knowledge. Scan physical documents. Utilize an AI tool like Claude Code, pasting prompts from the prompts/ directory (e.g., 01-tree-expansion.md, 02-cross-reference-audit.md). A comprehensive walkthrough is available in workflows/getting-started.md. No specific non-default software prerequisites are listed beyond the AI tool and markdown editor.

Highlighted Details

• Includes 12 autoresearch prompts specifically designed for autonomous loops, each detailing Goal, Metric, Direction, Verify conditions, Guard rails, Iterations, and Protocol.
• Provides a 19-file Obsidian vault starter kit (vault-template/) with core files and templates for various genealogical data types, using plain markdown for broad compatibility.
• Features 24 country and region-specific archive guides (archives/) detailing record locations, access costs, and AI tool compatibility.
• Offers 9 methodology documents (reference/) covering essential concepts like confidence tiers, source hierarchy, DNA interpretation guardrails, and AI capabilities assessment.
• Employs a philosophy of structured autonomous research with mechanical verification, probabilistic confidence, and logging of negative search results.

Maintenance & Community

Contributions are welcomed via pull requests for prompts, workflows, or archive guides. Examples should use placeholder names. No specific community links (e.g., Discord, Slack) or details on core maintainers are provided in the README.

Licensing & Compatibility

The project is released under the MIT License. This license generally permits commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

The project is explicitly "Built for Claude Code" but designed to be adaptable. Its effectiveness relies on the user's ability to scan physical documents and the capabilities of the chosen AI tool. The accuracy and completeness of the research are inherently dependent on the availability and quality of online genealogical sources.