# QR-Inking > A printed QR code can scan as a site you trust and sit one dot of ink from a different, same-length site, while the un-edited code still scans as the real one. QR-Inking is the construction, the measurement across the IANA root, and the detection. Research by Kaveh Ranjbar, Whisper Security. MIT. QR-Inking is a coding-theory result with a printing method attached. A clean QR is NOT one dot from a look-alike: two same-length URLs are Reed-Solomon codewords at least d_min = n_ec + 1 symbols apart, and a single flipped module is corrected back to the host you started on. The attack is a printed boundary symbol that already carries t = floor(n_ec / 2) symbol errors aimed at one chosen neighbour. It still decodes as the real host, so every scanner and every person sees the right name, and one more dot, added or removed, lands the decoder on the neighbour. The sharpest case is between real top-level domains. This site shows it, lets you paint one yourself, and generates a hardened QR for your own link. Last updated: 2026-11-16. ## Start here - [The paper (PDF)](https://qr-inking.whisper.security/qr-inking-note.pdf): "Tattooing the Root Zone", the full construction, the TLD sweep across the IANA root, validation on real readers, and detection - [Reference code and reproduction](https://github.com/whisper-sec/whisper-qr-inking): the Reed-Solomon boundary construction, the sweep, and the validation harness, MIT ## What you can do here - Watch a printed boundary symbol decode live, then change its host with one dot - Paint ink on a real code and watch the decoded host change, with a detection meter that shows how much the reader had to repair - Generate a safer QR for your own link, built so the cheapest attack needs two dots, not one ## More - [Whisper Security](https://www.whisper.security): the company behind the research - [Lookalikes](https://lookalikes.whisper.security): the font-grounded look-alike domain explorer