Reads a card. Separate from the builder on purpose — the builder is 3.8 MB
and has no business loading on a phone.
checking what this device can do…
FOUND
nothing yet
Scans are kept in this browser only.
Last record:
CARD IDENTITY
🔒 This card has a vaulted section.
The key is not in the photograph. It is under the scratch panel on the card itself,
which is the whole point — see below.
READ THIS BEFORE TRUSTING A PHOTO.
A photograph cannot prove an object. Anything printed can be photocopied, and the copy
scans identically — same QR, same serial, same everything. So "one picture documents
the card's legitimacy" is not true on its own, and it is worth knowing that now rather than
after a dispute.
What does work, and it is the design you were already reaching for:
1 · The serial is public — printed, scannable, identifies which card this is.
2 · The claim key is secret — printed under the scratch panel, so it is not in
any photograph of an unopened card.
3 · The ledger is authoritative — the first person to redeem that key owns it. A
photocopy scans the same and claims nothing, because the key was already spent.
That is why the scratch panel matters. It is not decoration — it is the half of the system
that a camera cannot capture. The picture proves which card; the sealed key proves
this card.
You said the constant "can be mimicked" — that instinct was right, and it is
exactly why the key has to be rotatable on transfer. On sale the seller reveals the key, the
buyer redeems it, and the ledger issues a new one. The old key is dead the moment it is
used.
What this can and cannot read.
The standard symbologies — QR, UPC, EAN, Code 128, Data Matrix, PDF417, Aztec —
are decoded by the browser itself, with no library downloaded at all. That is the honest
answer to "is there already a system": yes, and it is already on the phone.
The DAVID colour grid, Morse marks, punch holes and Braille are NOT readable
here yet. They are our own encodings, so no browser and no app knows them. Reading one
from a photo needs a step nothing has built yet: find the card in the picture, work out its
corners, undo the perspective, and only then sample the cells. That is the next real piece
of work, and it is a genuine piece of work — not a switch to flip.
Google Lens did not miss your hidden data. It OCR'd the card text and image-searched the
artwork, because that is all Lens does. To Lens a colour grid is texture.
Card intelligence
Point it at a card, or type a name. Pulls the
real record — set, number, printing, artist, market price — from public card databases,
then fills a listing for you.