A New Free Research Tool for Mid-Atlantic and Southern Genealogy
I've built a free research companion called Mid Atlantic and Southern Kin Guide — a custom GPT designed to help genealogists work through the kinds of records I spend most of my time indexing: wills, inventories and appraisals, tax lists, deeds, probate files, court minutes, church records, Freedmen's Bureau records, pensions, and local newspapers.
It's available to anyone with a ChatGPT account, free to use, and ready now.
What it's built to do
Most general-purpose AI tools approach genealogy the way they approach everything else — quickly, confidently, and without much regard for evidence. That's a problem for the kind of research I care about, where the difference between a documented fact, a reasonable inference, an oral tradition, and an assumption matters enormously.
Mid Atlantic and Southern Kin Guide is built to slow that down. It's designed to:
Analyze and interpret records, not just summarize them. It looks for clues, flags reliability concerns, and suggests what to research next.
Help you abstract and cite records in genealogy-standard styles.
Build the working documents researchers actually use — timelines, family group sheets, research logs, FAN-network notes, and simple family trees.
Explain historical context when it matters: county formation, boundary changes, slave laws, Civil War and Reconstruction-era realities, migration patterns, and local customs.
Treat family stories and oral history as evidence worth testing, not dismissing — while honestly distinguishing between what's documented and what isn't.
Adapt to your skill level, whether you're untangling a copied online tree for the first time or applying advanced FAN-network strategies.
What it's particularly good for
It has special depth in:
Research into enslaved and free persons of color, enslaver households, and the communities around them — including the strategies that make this research possible: studying Black and white families together when historically connected, working across county and state lines, and combining ordinary records in creative ways.
The Mid-Atlantic and Southern states, with attention to local history, county formation, and the kinds of records each jurisdiction produced in each era.
Difficult records — confusing handwriting, unfamiliar terminology, partial or conflicting documents, and the records that survive when so many others didn't.
A few notes before you start
It's a tool, not an oracle. Even with careful instructions, AI can be wrong — about names, dates, places, and especially about specific people. Treat its work the way you'd treat a research assistant's notes: useful starting points that need your verification before they go into your tree or your reports.
It can't see your living relatives' information unless you share it, and conversations with it are private to you. As with any AI tool, I'd gently suggest not sharing sensitive information about living people.
And it's new. If you try it and something works beautifully — or something falls flat — I'd genuinely like to know. Feedback shapes the next version.