What does
HWT‑XXXX‑XXXX mean?
You saw a short code at the end of something — a post, an essay, a cover letter. Here's what it is, what it tells you, and what it doesn't.
Someone typed it.
You can watch.
An HWT code is a maker's mark for writing. The writer typed the piece into the HumanWroteThis editor, every keystroke was recorded, and the finished text was locked to a permanent code. Look the code up and you can watch the piece being written exactly as it happened — pauses, typos, deletions and all.
The text, exactly as certified.
The code is locked to one piece of writing at the moment it was certified. It can never be reused or moved to different text — if the words don't match the lookup, the lookup wins.
The writing, replayed.
Every keystroke was recorded in order as the piece was written — no paste, no drag-and-drop, no dictation. The replay shows the draft taking shape the way it actually did.
The date, and the writer.
Each certificate shows when it was made and how long the writing took. Writers with an account can put their name on it; others certify anonymously. Either way, the record stands.
Look it up. It takes seconds.
The questions a
skeptic should ask.
A tool about honest writing should answer hard questions honestly. These are the real ones, including the one you're probably already thinking.
Couldn’t someone generate text with AI and just retype it by hand?
Yes. We'd rather tell you that than have you find it out in a comment section. A certificate doesn't claim the words couldn't have come from somewhere else — it documents the process: this exact text was typed, character by character, by a person at a keyboard. That's why we describe a certificate as evidence of process, not proof of no-AI. Nobody can prove a negative about AI — not us, not the detectors. What a writer can do is show their work. And the replay is worth watching before you judge: a real draft looks like writing — hesitations, backspaces, sentences torn up midstream. Transcription looks like transcription.
So is this an AI detector?
No — closer to the opposite. Detectors look at a finished text and guess, and they guess wrong often enough to hurt real people. HumanWroteThis doesn't guess about anything. It records how a piece of writing was made, while it's being made, and lets anyone watch. No score, no verdict, no accusation — just the record.
What exactly does a certificate claim?
One thing: the text tied to the code was typed by hand into our editor — no pasting, no drag-and-drop, no dictation, no automated input — and the keystrokes were recorded as it happened. It does not claim the writing is good, true, original, or untouched by AI elsewhere. The claim is modest on purpose. A modest claim is one we can actually stand behind.
What stops someone from pasting, or pointing a script at it?
Pasting and drag-and-drop are blocked in the editor, and input that arrives faster than human hands can type gets flagged. Is that airtight? No — no system is, and we won't pretend otherwise. Circumventing those controls violates our terms and voids the certificate. But the honest position is the one above: the certificate documents typing, the replay shows you the typing, and you get to judge what you see.
Can a code be faked, reused, or moved to different text?
A code is locked to one piece of writing and its recording when it's created — it can't be reassigned. If a code at the end of a post doesn't match what the lookup shows, the lookup is the source of truth. That's the whole point of checking: the code is just ink; the certificate is the record.
Does looking one up cost anything? Do I need an account?
No and no. Looking up a code is free and takes seconds. Making your own certificate is free too, and you don't need an account to do it — an account only matters if you want your name on your work.