The file with your name on it

Every one of them will mail you a copy of you.

Thirteen services on this page hold a record of what you did on them. Ask, and most will send it to you inside a week, as a zip. Below is what is actually in each one.

The contents are not the surprising part. The surprising part is that the file exists at all, ready to be assembled on request.

archive/ 13 services select one to open it

In the archive

    Recorded as

    Request yours ↗

    None of this leaked

    You agreed to all of it. That is what makes it useful to them.

    01 — It exists

    The export is the receipt

    A company can only send you an archive of something it already keeps. The request button is not a courtesy feature; it is a listing of inventory. Everything on this page is verifiable by anyone willing to click it and wait.

    02 — It is structured

    It arrives query-ready

    Not screenshots. Timestamped rows: what, when, with whom, from where. Tinder counts how many times you opened the app, by day. Snapchat keeps location history. LinkedIn keeps every job you applied for and every time you logged in.

    03 — It is being used

    Data is collected to be used

    Nobody stores a per-day count of app opens with no plan for it. Storage costs money and retention carries legal risk; both are accepted because the record earns its keep. You do not have to guess at motive to see that.

    So the asymmetry was never about what they know.
    It is that they can ask it questions, and you got a zip file.

    Get yours

    Thirteen requests, one afternoon, then a wait.

    Every link goes to that service's own data-request page. Start them all in one sitting — the archives arrive on their own schedule and several take weeks.

    Formats and delivery times as Holmes has them recorded. Services change both without notice.
    ServiceFormatTypical waitRequest

    Ask for JSON wherever a service offers a choice — Meta's HTML option is styled for reading and much harder to parse. TikTok deletes the archive four days after it is ready, so watch for that one.

    Then what

    A zip you can’t read is not much better than no zip.

    Thirteen archives is tens of gigabytes across a dozen incompatible schemas, searchable only by filename. Holmes is the other half: it reads all of it on your Mac, into one dated record you can ask questions of.

    See what Holmes does with it Read the source

    What it adds to the archives

    • Photos and documents — the folders on your own disk, which no service will mail you
    • Apple Health, MyChart and bloodwork — parsed, not summarized
    • iMessage — already on your Mac, no request needed
    • One dated timeline — every source normalized onto a single spine
    • A citation on every claim — back to the file and the line it came from

    Holmes is open source and runs locally. Reading the archives takes a language model, and you choose which one — including a local one, in which case nothing leaves the machine at all.