From factory to shelf, one scan at a time.
Idanta photographs each label as it is made, turns its random fiber and ink pattern into a fingerprint, and lets anyone match that fingerprint later with a phone. Nothing is added to the product, so there is nothing to copy.
Capture the label’s microstructure.
A camera on the line reads each label. An edge processor beside the line extracts the features and stores a compact fingerprint in about 100 ms, so the line never waits.
Bind every unit to its route.
When cases leave the plant, a scan links each fingerprint to the distributor, date and invoice. Data flows to your ERP, manufacturing systems, WMS and TMS, and other systems as needed.
Match with a phone, anywhere.
An ordinary 20 MP phone camera captures the label. The match result comes back with the unit’s identity and history.
Use the evidence.
Diversion investigations, counterfeit takedowns, returns checks and recalls all run on the same identity. One enrollment, many uses.
Three ways to enroll. Pick the one that fits your plant.
Enrollment is the only step that touches your operations. Each option avoids changes to the label and keeps production running.
On the packaging line
Industrial camera and edge processor over the conveyor. Up to 1,200 units per minute. Best for high-volume SKUs.
Before labels are applied
Labels are read on the roll or sheet before they reach the line. No camera near the product.
At the label printer
Your label supplier enrolls during printing. Covers every plant and contract manufacturer that uses those labels.
For label makersWhat makes it work now.
Reading 20 micron features at line speed, then matching them on an ordinary phone, was not affordable until recently. Here is what makes it work.
Substrates proven today: paper, polymer film, cartonboard, flexible packaging and printed aluminum tins. See all substrates
Physics-based features
Principled feature extraction using a novel statistical physics-based approach built upon substrate-printing interaction.
Edge inference
Optimized vision models run on low-cost embedded AI modules. Sub-100 ms processing per unit, deployed per line.
Phone cameras
20 MP and better cameras are now standard, which makes last-mile verification possible without special readers.
Hybrid model
Classical features combined with a convolutional network, for reliability and explainable results in uncontrolled light.
Built for wear
Tolerates smudges, dust, notches, discoloration, heat and humidity, and up to 30% damage to the read area.
Technical questions
What exactly does Idanta read on a label?
The random microstructure created when the label is made: paper fiber arrangement, ink bleed along print edges and surface texture, at a scale of about 20 microns. These features exist on every label and cannot be controlled or reproduced by any printer.
Can a counterfeiter copy the fingerprint?
Copying the artwork does not copy the microstructure. A reprinted or photocopied label has its own random fibers and ink bleed, so it produces a different fingerprint and fails verification.
What happens if the label gets scuffed or dirty?
The model is built for real supply chains. It tolerates handling marks, smudging, dust, discoloration, temperature and humidity changes, and up to 30% of the read area being damaged.
Which materials does Idanta work on?
Paper labels, polymer film labels, cartonboard (including GC1 and GD2 grades), flexible packaging and printed aluminum tins are proven today. Bare metal parts and new materials are assessed case by case by our engineering team before you commit. See all substrates.
Do we need special hardware for verification?
No. Verification uses an ordinary smartphone with a 20 MP or better camera. Enrollment on a high-speed line uses an industrial camera and a compact edge processor that we supply and install.
Does the camera on the line need to be validated like other equipment?
The enrollment camera is passive: it does not touch the product, make decisions on the line or reject units. For regulated industries, labels can be enrolled at the label printer instead, before they reach your plant.