The Data

Read the bean,
not the label.

Every lot on uncrack is decoded before it reaches the auction, so you bid on evidence, not impressions.

Objective

No palate, no bias. One instrument reads every lot exactly the same way.

Comparable

Because each lot is measured identically, you can line them up side by side.

Before you bid

Every number is known before the auction opens, no roasting, no guesswork.

From bean to numbers
STEP 1
NIR scan

A non-destructive near-infrared scan reads a 200 g green sample 96 times, a full spectral fingerprint, with no roasting and no sample lost.

STEP 2
16 components

EGGSTONE resolves that spectrum into 16 quantified chemical components, sugars, acids, lipids, protein, alkaloids and moisture.

STEP 3
Predicted profile

The engine maps the chemistry to a predicted profile: five Taste Potential axes and quality indices like uniformity, with bitterness handled separately as a roast-dependent potential.

Powered by data and intelligence

See what a palate can't.

A single scan holds about 156,000 raw signals, far more than any palate, or any cupping form, could ever register. The hard part isn't reading them; it's knowing which few actually predict the cup. That's what the Uncrack Engine does: it finds the meaningful patterns and turns them into the components, taste axes and quality signals you act on.

156,000features
raw signals in a single NIR scan
16named
components we surface from them
5taste axes
statistically derived from the spectrum
96×scans
per lot, averaged for stability
One lot's NIR spectrum · significant features
1100nm2500nm
Highlighted points are signals the model found meaningful, only some map to a named component.
How buyers use it
For roasters

Know body, sweetness and roast difficulty up front, and dial in faster, with fewer wasted batches.

For traders

Compare every lot on one objective scale, and see what actually justifies a premium.

For QC & buyers

Screen clean cup, uniformity and storage stability before a single bag hits the table.

For the curious · the full chemistrynormalized 0–1 · range across our library

Optional reading. These are the 16 measured components behind every signal above, grouped by what they actually do in the cup.

Acidity & Brightness
3 components

The bright, fresh, fruit-like acidity that lifts and structures the cup.

Organic acidsBright, fresh, fruit-like acidity that balances the cup0.19–0.56
Citric acidsLively citrus brightness; strongest in lighter roasts0.59–0.88
Malic acidsApple, pear and stone-fruit acidity; most vivid in light roasts0.26–0.65
Aroma & Flavor Complexity
3 components

Maillard and pyrolysis precursors that build aroma, depth and flavor nuance.

Amino acidsMaillard and Strecker precursor, nutty, chocolate and umami depth0.44–0.67
DiterpeneFruity and malty aromatics; adds complexity to flavor and texture0.51–0.97
TrigonellineDegrades to niacin and roast aromatics; lends sweetness and savory depth0.09–0.46
Sweetness Precursors
4 components

Sugars that caramelize on roast into sweetness and browning aromatics.

SucrosePrimary sweetness; fuels caramelization for caramel-like roast flavor0.45–0.81
GlucoseClean mild sweetness; caramelizes to caramel, nutty and chocolate tones0.46–0.92
FructoseCrisp fruity sweetness; converts to berry and tropical-fruit aromatics0.33–0.64
Mono sugarPowers Maillard and caramelization, roasted, deep-sweet, burnt-sugar notes0.48–0.87
Mouthfeel & Body
2 components

Weight, viscosity and creamy texture you feel in the cup.

LipidsBody, aftertaste and creamy, nutty texture; carries aroma0.24–0.53
Poly sugarBuilds body and umami, weight and texture in the cup0.12–0.29
Bitterness
3 components

Pleasant bitterness and chocolate-cocoa depth, plus the smoky structure that develops on roast.

CaffeineStimulant and perceived strength; a minor contributor to bitterness0.17–0.59
Caffeine subTheobromine and relatives, chocolate, cocoa and smooth umami0.42–0.67
Chlorogenic acidsBright acidity when light; smoky bitterness as the roast develops0.42–0.78
Stability
1 components

Storage stability and how reactively the lot develops on the roaster.

Water activityStorage stability and roast reactivity; balanced moisture = even development0.28–0.59
Reading a lot page
Taste Potential radar

Five predicted axes, acidity, aroma, flavor, sweetness and mouthfeel (0–10). The shape tells you where a lot leans before you ever cup it. Bitterness is shown separately, as a roast-dependent potential.

16 components

The full chemical bar chart, sugars, acids, lipids, protein, alkaloids and moisture, the evidence behind the profile.

Compare overlay

Pick the auction average or a peer lot and the comparison line overlays every chart — a + means this lot is stronger on that component.

NIR fingerprint

All 16 components folded into one radial shape — a lot's chemical signature at a glance.

Example · Taste Potential
AcidityAromaFlavorSweetnessMouthfeel
BOL7 · Alasitas Geisha
See it on a live lot →Browse auctions