About Our Clarke
Our Clarke collects the history of Clarke County, Virginia, and publishes the evidence alongside it — including the places where the evidence runs out.
What this collects
Stories about places in Clarke County: how they got their names, who lived and worked in them, what stood on a corner before the building that is there now. Around each story sits the machinery that makes it checkable — the people and places it names, the statements it rests on, the sources behind those statements, a chronology, recordings, and the questions nobody has answered yet.
Right now that is 53 stories, 690 people and places, 1459 statements, 764 sources and 82 chronology records. It is a small sample of a county's history and it is not trying to look like more.
How the pieces connect
A story is the way in. It reads as prose, but every paragraph is anchored to one or more statements, and every statement names the source it came from and the exact passage within it. Open any evidence drawer in a story and you drop straight through the narrative to the document underneath.
People and places are the other axis. The same statement that supports a sentence in a story also appears on the page for the person or building it concerns, so you can arrive from either direction. Connections between them are statements too, assessed on their own evidence: a relationship recorded in a later account is not proof that two people met.
The chronology keeps a distinction most timelines quietly drop — between the date something happened, the date of the document that records it, and the date a book about it was published. The archive holds pictures and recordings, and is honest about the ones we cannot show. Unresolved questions are published at the foot of each story rather than tidied away, because the gaps are part of the record.
How evidence status works
Every statement carries one of three labels, and each is judged on its own evidence. Nothing is averaged, scored, or voted on.
The named source supports this narrow statement. It does not mean anyone inspected an original document, or that a historian confirmed it.
The source hedges, so we hedge. The uncertainty in the original survives into what you read.
Somebody told this story. We cannot establish that it is true, and we say so rather than dropping it.
Two more things the labels do not do. They do not describe how interesting a statement is, and a well-documented account can be "supported" while the thing it describes stays unconfirmed — the label attaches to what the source shows, not to the past itself. The evidence page works one statement all the way down if you want to see it.
Who is behind this
Our Clarke is run by Michael Oak. Mike has lived in Clarke County for ten years. He married his wife, Jen, in Berryville on December 23, 2016. Both have always been passionate about history, and they were drawn to Berryville for its small-town charm and quiet rural setting.
This site is Mike's side hobby. He uses Claude Code, an AI tool, to research the history of Clarke County and to bring forward past articles by talented writers, and historical photographs from professional photographers, home snapshots and other archives. Every article and photograph keeps its author's or photographer's credit.
No money. No money has been solicited for this site, and advertisements are not allowed.
Our Clarke is independent. It is not affiliated with, endorsed by, or speaking for Clarke County, the Town of Berryville, the Clarke County Historical Association, any museum, library, university, church or business named on this site. Where an institution appears here, it appears as a subject or as the publisher of a source — never as a sponsor.
How this site uses AI
An honest note about the stories. As of October 10, 2026, most of the stories on this site were drafted with AI, and many read that way. We have done our best to filter out the worst of it, but these stories deserve a human telling, with a human voice, heart and mind. Our goal is for professional and hobbyist historians to write them in their own voice and from their own perspective. If you would like to write one, get in touch.
What the AI does here. It searches and reads digitized newspapers, books and records, transcribes what it reads, checks each statement against its source, and drafts text. Every statement links to the source it rests on, and every page says what has not been checked.
What it does not do. It does not contact anyone; Mike does. It does not name a living person without that person's consent. It does not invent sources, quotations or photographs, and it does not present its own drafting as a historian's work.
Sources and methods
Who did the reading. The research and source-checking were done with the help of Claude, an AI system. No independent historian has reviewed this work, no original document has been inspected in an archive for it, and nothing here claims an original discovery.
How a record reaches this site. Every record is checked against its source and given an evidence assessment before it can be published. A record appears here only when a release manifest names it explicitly, and only the fields chosen for publication are sent to your browser.
Three decisions are kept apart. Evidence assessment is how well a source supports a statement. Research state is how far the work has got. Release is whether it belongs on a public page. A well-evidenced fact can stay unpublished because its image rights are unclear, and a clearly attributed piece of lore can be published while remaining unconfirmed.
Photographs. An image needs an inspected file and an item-level rights basis before it can be displayed; a source link and a copyright label are not a license to reproduce. The photograph archive explains the gate in full.
The drawings
The drawings at the top of these pages are ours. The land is traced from USGS elevation data along the line of latitude through Berryville, with the height exaggerated so the valley and the Blue Ridge read at a glance. The county map is drawn from U.S. Census boundaries.
What is deliberately missing
No map drawn from the records, because they carry no coordinates and no settled historical boundaries, and a dot on a map reads as fact whatever the caption says. A historical map sheet in the archive is a document, not a layer, and nothing here is positioned from one. No photographs that are not cleared item by item. No accounts, no analytics and no tracking. The one thing this site does send anywhere is what you type into the submission form, which goes to a private review queue and is never published automatically. The colored panels on story pages are original typographic designs; they are not documents, maps, photographs or reconstructions, and each one says so.
This sample is not a claim that the county's history begins in 1688, or that anything absent from it did not happen. An empty page here is a statement about this project, never about the past.
Using this
The released collection is available as a single JSON file, with the same fields the pages use and nothing that is withheld from them. Your reading list is stored in your own browser and goes nowhere else. Source links open other people's websites.