How NFC tags authenticate products

NFC tags for product authentication are small radio tags fitted into a product, which a phone reads with a tap. Near Field Communication runs at 13.56 MHz over a few centimeters: the NFC Forum gives a typical range of up to 2 cm. A tag on its own only identifies. It authenticates when each tap gives the phone something a copy cannot produce.¹

An NFC authentication tag is a tag that proves it is the one the brand fitted, every time it is read. It has no battery: NFC carries power as well as data, which the NFC Forum notes allows battery-less operation, so the phone's field powers the tag for the moment of the tap.¹

The proof comes from a secret key held inside the tag. On each read the tag uses the key to compute a short code, and a server that knows the key checks it. Cryptographers call this kind of code a message authentication code: the US National Institute of Standards and Technology (NIST) defines it as a value computed with a secret key and used to establish that data is authentic. The key is never part of what the phone reads, so reading the tag, or copying what it sent, gives nobody the means to make the next code.³

Most NFC tags are plain. They store a link and other data in a standard record, which the NFC Forum calls an NFC Data Exchange Format (NDEF) payload, and a phone reads it and acts on it. Locking a plain tag against rewriting stops someone changing it, not someone copying it. A link copied from a plain tag is the same link, and it reads the same way from a web page or from a second tag.¹

A secure tag adds cryptography to the read itself. It still holds a link, but each tap adds a fresh code computed with the tag's key, and a count of how many times the tag has been read. Secure tags follow the same contactless standards as plain ones, so ordinary phones read them in the same way, and the difference only shows when the code reaches the server.

Phones read NFC tags in that standard record format and act on what it holds. For a web link, that means opening it or offering to: the NFC Forum describes a phone showing a dialog that confirms the URL to visit. Android's documentation says that from Android 16 a tag holding a web link opens it directly, and that Android 17 first shows an "open link" notification for the user to tap.^1,2

With a secure tag, the link carries the code. The phone opens the link, the code travels with it to the brand's server, and the server checks it and returns the result. All of this happens inside an ordinary read of an ordinary link, which is why a tap can open a page in the phone's browser rather than a dedicated app.

Security standards have a name for passing off a captured message as a fresh one: NIST calls it a replay attack, in which someone replays previously captured messages to pass as the genuine party. A secure tag defends against it by making every code single-use. The server remembers which codes, and which read counts, it has already seen for each tag, and turns away any it has seen before or that arrive out of order.⁴

That is what defeats the easy copies. A screenshot of the page, a forwarded link, or the URL written onto a cheap plain tag all carry a code the server has already seen. The code itself is genuine; it is simply used up. A counterfeiter would need the secret key to make the next one, and the key stays in the tag.

A plain tag can be, in the sense that matters: its stored link and data can be read and written onto another tag. A secure tag is built to resist that, because the key that makes its codes stays inside the tag.

What a secure tag changes is what a copy can do. A copy of the link is not enough to pass for the original, it cannot make a new code without the key, and a replayed code shows up as one.

A tag only proves something while it stays with the item, so placement is decided before the tag goes in. The rule is the same in every category: fit the tag where it is protected, usually out of sight, and hard to move to another object without damage.

Each read of a Vera Tag produces a signed, single-use response, and the secret that signs it stays in the tag, so reading a tag's identifier does not give anyone the ability to verify. A copied or screenshotted link is not shown as verified, a replayed tap is flagged on the item's certificate for the brand and for whoever holds the item, and the brand can see where suspicious scans happen.

A tap opens the certificate in the phone's own browser, with nothing to install, and the tag needs no battery. Vera encodes every tag before it ships and can bind a box of tags to a product first, so each item carries its identity from the factory, before it is sold, and keeps it through every owner after. A damaged tag is replaced and the certificate moves to the new one with its whole history.

• Does every tap produce a new code? A tag that returns the same link each time identifies the item but cannot prove it.
• Does the server reject a code it has seen before? A copied link should never show as verified.
• Where does the key live? It should stay inside the tag, never in the link or on the packaging.
• Can a customer check it with the phone they already have? Every extra step, such as an app to install, loses people.
• Is it fitted as part of the item, somewhere it cannot be moved to another item without damage?
• What happens when a tag is damaged? The record should survive the repair.