Draw ground anywhere on Earth and get a first-pass read on it — built from the same trained evidence as our national models, and straight with you about where it is strongest.
Sign up freeMost of the world has never been systematically assessed for prospectivity. The national geochemical survey programs that make a certified model possible exist in a handful of countries; everywhere else, an explorer opening a new jurisdiction starts close to blank. The global model is built for exactly that moment. It reads your ground from evidence available everywhere on Earth, scored by models trained on the deposits and survey chemistry of three of the best-mapped mineral provinces in the world — so you can shortlist ground worth a field team before anyone flies.
minerals you can run worldwide
countries behind the training (45 · 51 · 37 minerals)
minimum billed area, priced in USD
The global model is trained on up to 1.5 million geochemical field samples per mineral from the national survey programs of Australia, the United States and Canada — the same foundation behind our certified models. It applies that training through evidence that exists everywhere on Earth: satellite spectral response, radar, terrain, and mapped rock families. That combination is what lets a model learned in the Yilgarn or the Basin and Range say something useful about ground it has never seen.
Learned from the survey chemistry and known deposits of three of the most thoroughly mapped mineral provinces on Earth — not from theory, and not from a global average.
Satellite spectral and radar response, terrain, and mapped rock families — consistent worldwide, so the same reading applies in Mauritania as in Mongolia.
Your ground is read against the rock families it actually belongs to, so a granite is compared with granites rather than with whatever happened to be nearby.
A global run returns a full prospectivity map of your area and the context to act on it — read together, they tell you which ground deserves the next field visit.
Every readable cell in your area scored for the minerals you chose, so the strongest ground stands out at a glance.
Cell by cell, how closely your ground matches the terrain the model was trained on — the difference between a lead worth chasing and one worth sampling first.
The geology under the signal, so you can see which units are driving the strong ground.
Exactly how much of your area could be read at all — stated up front, before you pay.
In Australia, the United States and Canada we hold dense national geochemistry, so a model can be tested the hard way — hide the known deposits, rebuild without them, and check whether it still finds them. That earned test is what lets those models rank drilling targets. Outside those countries that test ground does not exist, so a global run gives you the prospectivity map, the geology behind it and the confidence to read it — and leaves the drill-target ranking to the countries where that test was earned. It is the difference between “this ground is worth a closer look” and “drill here”, and knowing which one you are holding is what makes either useful.
| Certified — Australia · USA · Canada | Global — everywhere else | |
|---|---|---|
| What a run returns | Ranked drilling targets with the evidence behind each | A prospectivity map with the evidence and confidence behind every cell |
| How it was tested | Held-out known deposits, kept spatially separate from training ground | Trained and tested in the certified countries, then applied to new ground |
| Confidence read | How closely the ground matches known mineralised systems | How closely the terrain matches the model's training experience |
| Best used for | Drill targeting and program design | Ground selection, licence and acquisition decisions, portfolio triage |
The model reads the surface. Bare and sparsely vegetated terrain — the arid and semi-arid belts where much of the world's frontier exploration happens — is where it reads best.
Densely vegetated and water-covered ground sits outside the model's valid domain. We exclude it rather than guess at it, and we tell you what share of your area that is before you pay.
Where your geology looks unlike anything in the training regions, the confidence layer says so cell by cell instead of painting the map green.
A mineral is offered globally only where the model earned it. Where it did not, the product refuses it rather than serving a number we do not stand behind.
The global model is a starting point, not a ceiling. Where you already hold geochemistry — historic assays, soil surveys, drill data — we can fold it in and calibrate to your region, which is how a global read becomes certified-grade modelling for your ground. Every project strengthens with local samples, and we will tell you honestly if yours will not move the needle.
Those three countries have dense public geochemical surveys, so a model can be validated on known deposits it never saw in training — and that earned test is what lets it rank drilling targets. Outside those countries that validation ground does not exist, so a global run gives you the prospectivity map, the geology driving it and a confidence read on every cell — everything except the drill-target ranking, which we publish only where it was earned.
Within its stated purpose, yes — mineral systems leave comparable surface footprints in comparable rocks, which is why the model reads your ground against the rock families it belongs to. But you should not take it on faith: every map carries a confidence layer showing, cell by cell, how closely your terrain matches what the model was trained on. Where that match is weak, treat the result as a prompt to sample rather than a conclusion.
Most of the catalogue, but not all of it. A mineral is offered globally only where it cleared our release gates; where it did not, the product declines to run it rather than returning a number we do not stand behind. The coverage table on the home page shows exactly which minerals are available where.
They are excluded. The model reads exposed ground, so densely vegetated and water-covered terrain falls outside its valid domain and is masked out rather than guessed at. Before you pay, the app shows what share of your area can actually be read — so a heavily vegetated block never comes as a surprise after the fact.
The same per-area pricing as anywhere else — you pay for the ground you scan, with the price shown before you pay and a 25 km² minimum billed area. Runs outside Australia, the United States and Canada are priced in US dollars.
That is exactly the path. Where you hold local geochemistry we can incorporate it and calibrate the model to your ground, which is how a global read becomes certified-grade modelling. Talk to us about what data you have and we will tell you honestly what it would buy you.
Draw an area, pick a commodity, and explore the ranked targets and the evidence behind each.
Sign up free