← Back to the tree

Import a research file

Bring an existing family tree into MyAncestors in one batch. Every imported person lands in the review queue — nothing goes live until an admin approves it.

1. Format
Hand the prompt below to your own AI (ChatGPT, Claude, …) along with your notes. It returns a formatted file.
2. Import
On the home page, open Family tools → Import a research file, then paste the text or upload the file.
3. Review
The people appear as one batch in the admin review queue with their facts, sources, grades, and any flags — approve or reject.

The prompt for your AI

Copy this, paste it to your AI, and add your research where it says so. The AI does the formatting; the import itself uses no AI.

Download prompt (.txt)
Loading…

The format, field by field

One block per person, each starting with a line that is exactly --- PERSON ---, then Label: value lines. Omit any label you have no value for. Names in the kin fields are plain names — a reviewer links them to real people afterward.

LabelWhat it is
Name:Required. The person's full name.
Nickname:A nickname, if any.
Maiden name:Surname at birth (for someone recorded under a married name).
Sex:M or F — only if the source states it. Never guessed.
Born: / Died:date | place — either side may be blank. Partial dates like ~1705 are fine.
Parents: Spouses: Children: Siblings:Comma-separated names.
Marriage:spouse | status | date | place. Status: married / divorced / widowed / separated / partners / unmarried / unknown. One line per marriage.
Confidence:A (documented) / B (corroborated) / C (compiled or tradition). Use weak if the source flags the link as unproven.
Flag:type | note — a caution shown in the tree. type: warning, structural, or general.
Note:Free text. Put burial, baptism, and occupation here for now.
Source:A citation. Repeat for more than one.
Unverified:Something that still needs verification — never a stated fact.

A worked example

Download example (.txt)
Loading…

Confidence tiers become each person's reliability grade (A/B/C), and a weak tier adds an “unverified” flag. Warnings and structural questions become flags on the person, so nothing questionable is silently committed. Facts a field can't hold yet (burial, occupation) are kept in the person's notes.