Text-Hygiene Engine Checks and cleans what you publish: invisible characters, hidden metadata, origin labels
The Text-Hygiene Engine checks and cleans what you publish: it removes invisible characters from text, and hidden data (metadata and C2PA origin labels) from documents and images. It does not claim to remove watermarks hidden in word statistics: that cannot be verified, and it says so. $99. Not on sale yet: join the waiting list.
What it does: Check and clean the text, documents and images you publish of invisible characters, metadata and origin labels (C2PA), in the formats listed below, and know exactly which layer you cannot verify.
Who it is for
Anyone publishing AI-assisted text or documents who has to show what was cleaned, file by file.
Not for: Anyone hoping to get past an AI detector. That is not a feature and not a promise.
What it solves
- Invisible characters and stray metadata travel with the text and nobody sees them. A check run on every publication finds and removes what can be verified. 0 zero-width characters and 0 byte-order marks on every publication since 2026.
- Cleaning tools claim more than they can prove. The layers that cannot be verified are named as such, in writing. Declared unverifiable, in writing, since 2026.
- A clean file today says nothing about the next one. A per-publication checklist and an audit log per page. Kept with each article's production record.
Price and access
Files and their updates are delivered in English.
Waiting list: Text-Hygiene Engine
No ship date is set: a system goes on sale once the points under “What we have not verified yet” that hold it back are closed and a live purchase of it has been tested. You get one dated note that day, with the price and the proof attached. No other emails, no countdown.
On sale now: The Domain-Only Website Kit, $129.
What is inside
The engine, written by us and under the MIT licence (you may reuse and change the code freely): the Python version covers text, documents and images, and the JavaScript (Node) version covers text only. With it come the protocol that separates the verifiable layers from the statistical one, the per-publication checklist we run, the metadata purge scripts for PDF and images, and the written list of what cannot be verified and why.
New to the problem? Read what a zero-width space is.
What we measured on the sites we manage
Not on sale yet: this product goes on sale once the points under “What we have not verified yet” that hold it back are closed and a live purchase of it has been tested.
Measured with: the engine's own audit log per publication, kept with each article's production record; PDF metadata purge script in the toolkit.
The proof is kept in one dated audit folder and shared on request; every figure is also on our figures page.
What we have not verified yet (unverified)
- Value against the free tools: Their base function is free. What is sold is the full process, the log and the list; whether buyers pay for that is what the first sales will tell.
Text-Hygiene Engine: questions
Does it hide text from AI detectors?
No, and no page here says it does. It removes the verifiable marks (invisible characters, metadata, provenance) and declares the statistical layer unverifiable.
What formats?
Plain text and Markdown with either version; DOCX, PDF and common image formats with the Python version only. From the command line, with a per-file log.