An uncopyable QR code is a promise nobody can keep - anything printed can be photographed. So we built the opposite: a copy-evident code whose fine structure is destroyed the moment it is reproduced, and a network that watches how every identity behaves. Verified on a standard smartphone. No chips, no readers, no app.
Identification answers which product this claims to be. Authentication answers whether this item is real. A serial number does the first and not the second - it can be read off one genuine product and printed onto a thousand fakes, each carrying a number that looks perfectly valid. Every counterfeit strategy that relies on numbers alone fails at exactly this point.
So the question is not how to make a better number. It is how to make the physical thing carrying that number impossible to reproduce unnoticed.
Inside every mark sits a crypto-optical structure - detail fine enough to survive your production press, but not a photocopier, a photograph or a re-print. Reproduction smears it. The check looks for exactly that.
Illustration. Both codes look identical at a glance - the difference lives in detail too fine to survive reproduction. Zoom in at 4× to see it.
Assume someone defeats the print. They still cannot fake where the code has been. Two layers have to hold at once - and the second one gets stronger the more product you protect.
Fine structures inside the mark are cryptographically derived and physically fragile. They survive industrial printing but degrade under reproduction - so the mark itself carries the evidence, no reference sample required.
Every scan writes a time and a place. One identity in two countries at once, a unit surfacing far outside its allocated route, a burst of scans on a code that should have sold once - each is invisible in a single scan and obvious across the network.
The mark is printed into the packaging you already produce, or applied as a label. No new hardware on the line.
Each unit is bound to a unique, cryptographically derived identity - not a number you could guess or increment.
The identity is registered, so it can be verified and every scan it ever receives is written to its history.
Anyone points a phone. The structure is checked, the behaviour is checked, the answer comes back in seconds.
Every approach trades something away - cost per unit, a reader in someone's hand, a separate part on the line, or the ability to tell a copy from an original at all. This is where each one lands.
| Approach | Hard to clone | Smartphone | No hardware | Low cost at scale | DPP-ready |
|---|---|---|---|---|---|
| Standard QR code | ✗ | ✓ | ✓ | ✓ | ~ |
| Serialised QR code | ✗ | ✓ | ✓ | ✓ | ✓ |
| NFC | ✓ | ~ | ✗ | ✗ | ~ |
| RFID | ✓ | ✗ | ✗ | ✗ | ~ |
| Hologram / security label | ~ | ✗ | ✓ | ~ | ✗ |
| Blockchain-only | ~ | ~ | ~ | ✗ | ~ |
| authentic.network | ✓ | ✓ | ✓ | ✓ | ✓ |
Comparison of technology approaches, not of individual vendors.
The technology is not a pilot. It runs in live rollouts across tools, lubricants, solar and healthcare - categories where a fake is a safety problem, not just lost revenue.
A standard QR code is a container for data. Anyone can photograph it, reprint it and the copy scans exactly like the original - because nothing about the code proves it is the original. The authentic.network mark adds a physical layer: fine structures inside the code survive industrial printing but degrade when the code is reproduced, so a copy can be told apart from the genuine item.
They can photograph and reprint anything. That is the point of the design: the copy is possible, but it is not identical. Reproduction destroys the fine structure the check looks for, so the copy announces itself as a copy. The honest claim is copy-evident, not uncopyable.
Serialization answers which product this claims to be. Authentication answers whether this item is real. A serial number can be read off a genuine product and printed onto a thousand fakes - each one carrying a number that looks perfectly valid. Identification is not authentication; the physical mark supplies the part the number cannot.
A hologram is a separate component: it has to be produced, sourced, applied and kept in sync with the product it protects, and judging it usually needs a trained eye or a reference. The authentic.network mark is printed into the packaging you already produce, needs no extra part on the line, and is judged by software rather than by the person holding it.
A printed structure whose fine detail is cryptographically derived and physically fragile. It survives your production press but not a photocopier, a photograph or a re-print - which is precisely what makes the difference between an original and a reproduction machine-readable.
Two layers. The printed structure shows whether this particular mark is a reproduction. On top of that, scan behaviour is analysed across time and geography - the same identity appearing in two places at once, or scanning far outside its allocated route, surfaces as an anomaly even when each individual scan looks fine.
No. Verification runs on a standard smartphone camera. There are no chips to embed, no readers to install and no app for your customer to download.
That is the design goal - the mark is applied as a label or printed directly into your current process rather than requiring a dedicated line. Print method, resolution and substrate do affect how the structure performs, so your setup is verified before rollout rather than assumed.
The technology provides the verifiable identity layer a Digital Product Passport needs, engineered for ESPR alignment. A passport is only as trustworthy as the code carrying it: if the data carrier can be copied onto a counterfeit, the passport travels with the fake. Specific obligations depend on your product group and its delegated act.
A 20-minute call is enough to map the mark to your substrate, your press and your volumes - and to tell you honestly where it fits and where it doesn't.