How it works

A molecular fingerprint for every product.

NIR Watchdog builds a molecular fingerprint of a product from a Near-Infrared spectrometer reading, using machine learning, and flags any deviation from the validated standard — current measured detection accuracy above 98%.

See it in action

Watch the platform work.

From field scan to verdict — a short demo of NIR Watchdog in action.

Training

From validated samples to a living model library.

01
Reference samples
Validated authentic material is collected as the ground truth.
02
NIR fingerprinting
Spectral signatures are captured across the material.
03
Model training
Deep learning and transfer learning build the model.
04
Model library
A dedicated model per material, per device.
Inspection

Scan in the field, get a verdict in seconds.

01
Field scan
Any portable NIR device paired with the native mobile app.
02
AI comparison
The reading is compared against the validated model.
03
Verdict in seconds
Authentic · deviating · anomalous.
Evidence on the XRP Ledger

An immutable audit trail for regulators and customers.

Every scan and model version is hashed, signed and timestamped — spectra, device ID, location and result — creating a tamper-proof record you can hand to auditors, regulators and customers.

Any material, any process

Works on any solid or liquid material, and integrates with the quality control you already run.

Solid or liquid

From tablets and powders to oils, water and beverages — the same platform, new models.

Above 98% accuracy

Measured detection accuracy across pharma, food, water and supplements.

Fits your QC

Slots into existing quality-control processes — no lab, no waiting.