More in Family history.
Transcribe an 1829 handwritten deed with Claude and check it against a manual copy
WorkedEvidence: Screenshots or logsPosted Last checked
Claude 3.5 Sonnet transcribed an 1829 deed almost perfectly for genealogist Nicole Dyer, but misread the key surname Peerson.
Beginner-friendlyNo logins, payments or coding needed. Low risk: paste the prompt and you should get something useful.
How to use this
- Tap Copy prompt.
- Open Claude.
- Paste it into the message box.
- Send it, then read the answer before acting on it.
New to this? What's a prompt?
A prompt is the message you send to the AI tool. Copy the one on this page, put your own details where it asks, and read the answer before you act on it.
Prompt
These are the person's own words, as shared in the source.
You are an expert genealogist and transcriptionist. Transcribe this deed and put anything you can’t read in square brackets with question marks or your best guess. Preserve the original line lengths in the doc.
Try this in:
Sends the Original prompt shown above. ChatGPT and Grok open with it filled in, and ChatGPT may send it straight away. Gemini and Claude open empty, so we copy the prompt for you to paste. Every button also copies it.
The source used Claude. These are general chat tools, so results may differ.
Opens in a new tab. We don't see what you send.
These links are only a shortcut. We're not affiliated with OpenAI, Google, Anthropic or xAI, and your results may not match the source.
Optional stronger version
Our rewrite with extra checks. The poster didn't test it.
Before you start, ask me 1–3 short questions about the details below and wait for my answers. Once I've answered, confirm the plan in one line and run the whole task start to finish. Goal: Transcribe a handwritten historical record exactly as written, flagging anything uncertain, so I can check it line by line. Questions to ask me first: - What kind of record is it, and from where and when? - Which names, places and dates do you already know from other records? - Do you want spelling kept exactly as written? Inputs (fill in or confirm): - Image of the record: [IMAGE] - Record type, place and date: [e.g., DEED / LINCOLN COUNTY, NC / 1829] - Known names and places: [NAMES] - My own transcription, if I have one: [TRANSCRIPTION] Rules: - Keep original spelling, capitalization, abbreviations and line breaks; don't modernize anything. - Put unreadable or uncertain words in [brackets?] with your best guess. - Mark superscript letters as written. - Treat names and dates as the highest risk and list each one separately for me to check. Checkpoint: Transcribe the first 10 lines and wait for me to check them before continuing. If you get stuck: If the image is too faint or cropped, tell me which lines and ask for a closer image. Verify before you finish: If I give you my transcription, list every difference between ours, line by line, without assuming either is right. Output: A line-by-line transcription, a list of uncertain words and a names-and-dates checklist. Done when: I have a transcription I've checked against the original image.
Customize the stronger version
Answer these questions and we'll put your answers into the stronger version. The original prompt above stays exactly as published.
Your file stays on your device. Nothing is uploaded to TriedPrompts.
Your prompt
Before you start, ask me 1–3 short questions about the details below and wait for my answers. Once I've answered, confirm the plan in one line and run the whole task start to finish. Goal: Transcribe a handwritten historical record exactly as written, flagging anything uncertain, so I can check it line by line. Questions to ask me first: - What kind of record is it, and from where and when? - Which names, places and dates do you already know from other records? - Do you want spelling kept exactly as written? Inputs (fill in or confirm): - Image of the record: [IMAGE] - Record type, place and date: [e.g., DEED / LINCOLN COUNTY, NC / 1829] - Known names and places: [NAMES] - My own transcription, if I have one: [TRANSCRIPTION] Rules: - Keep original spelling, capitalization, abbreviations and line breaks; don't modernize anything. - Put unreadable or uncertain words in [brackets?] with your best guess. - Mark superscript letters as written. - Treat names and dates as the highest risk and list each one separately for me to check. Checkpoint: Transcribe the first 10 lines and wait for me to check them before continuing. If you get stuck: If the image is too faint or cropped, tell me which lines and ask for a closer image. Verify before you finish: If I give you my transcription, list every difference between ours, line by line, without assuming either is right. Output: A line-by-line transcription, a list of uncertain words and a names-and-dates checklist. Done when: I have a transcription I've checked against the original image.
Scroll to see the whole prompt. Your answers show in green.
3 blanks are still empty: Record type, place and date; Known names and places; My own transcription, if I have one. They'll stay in brackets in your prompt, so the chat tool may ask you about them.
What you type stays in this browser tab. We don't send it or save it, and it's gone when you leave the page.
AI tool
Outcome
Genealogist Nicole Dyer of Family Locket tested Claude 3.5 Sonnet on an 1829 handwritten deed from Lincoln County, North Carolina, which Dyer had already transcribed by hand. Dyer gave it the image and asked it to transcribe the deed, mark unreadable words in brackets and keep the original line lengths for easier checking. Dyer called the result stunningly accurate; its main mistake was changing the grantor's surname from Peerson to Person. Dyer then gave Claude the manual transcription and asked it to list the differences. It found 15. In four, Claude was right and Dyer was wrong. In two, both were wrong: the original showed Pottses Creek and the neighbor's surname Altom. The rest were small errors such as capitalization and superscript letters.
Evidence: Screenshots or logsThe source shows screenshots, a transcript or the AI's output.
Source
- Posted
- by Nicole Dyer (Family Locket)
- Last checked
- Method:
- How we verify
https://familylocket.com/handwritten-text-recognition-by-claude-3-5-sonnet/
Context: This is a 2024 test with an older Claude model. It misread a key surname, and the author notes that models struggle with proper nouns and dates. A commenter had worse results on a different will. Always check against the original image.
More in Family history
Catch a wrong birthplace in a family tree with a NotebookLM Audio Overview
A NotebookLM Audio Overview made a genealogist notice that an ancestor's recorded birthplace didn't match the records.
WorkedNotebookLMEvidence: Described by posterBeginner-friendlyView CaseSearch 40+ sources on a brick wall ancestor's FAN Club with NotebookLM
NotebookLM searched 40+ sources at once with citations and helped Lisa Rex confirm a migration pattern from Ohio to Illinois.
WorkedNotebookLMEvidence: Described by posterBeginner-friendlyView CaseCompare four AIs transcribing a 1700s will found with FamilySearch full-text search
Dana Leeds tested Copilot, Gemini, ChatGPT and Claude on a 1700s will. Claude was most accurate, but each made spelling mistakes.
PartialClaude +3Evidence: Screenshots or logsView CaseTranslate a great-grandfather's Yiddish letters for a family archive with Gemini and Claude
A man used Gemini and Claude to translate Yiddish family letters from 1900 to 1926, checking for made-up text with native speakers.
WorkedGemini +1Evidence: Article or interviewView CaseTranscribe and date 150 old family letters for an archive with ChatGPT
A genealogy volunteer used ChatGPT to transcribe about 150 letters from 1885 to 1925, then date them from clues in the text.
WorkedChatGPTEvidence: Described by posterView CaseSeparate land descriptions in an 1836 dower record with ChatGPT
ChatGPT separated a 113-acre tract from a 43¾-acre dower in an 1836 record and linked the widow to a figure in medical history.
WorkedChatGPTEvidence: Described by posterView Case

