Explainable AI for
mineral prospectivity mapping.
Built on 13.8 million samples and the full public geoscience of every country we cover, MineDSS ranks the ground for 51 minerals — so your next drilling program goes where the evidence points. No black box: every target comes with the reasoning behind it, so you can see why it ranks where it does. And with our global model, you can now run ground anywhere on Earth.
Free to sign up — and no login needed to explore the live demos below.
See it on real ground.
Explore 11 real example runs across Australia, the USA and Canada — the prospectivity map, the per-target reasoning, the geology layers and the analogue-support signal. No login. Click any area to open it live.
Carlin, NV 
Sudbury, ON 
Cobar, NSW 
Morenci, AZ 
Val-d'Or, QC 
Olympic Dam, SA 
Mount Isa, QLD 
Mount Isa, QLD 
Kalgoorlie, WA 
Avebury, TAS 
Charters Towers, QLD & run your ground →
Real areas across three countries, real spread — from Sudbury and Cobar reading near-fully analogue-matched to frontier ground where we'd recommend a few field samples first. The model concentrates confidence where the evidence is; it doesn't paint everything green.
Every target comes with its reasoning.
A self-serve prospectivity engine that shows its work — the evidence behind every ranked target, the depth behind the signal, and an honest read on how far to trust it.
The reasoning, laid out
Every ranked target shows its evidence chain — host rock, geophysics, geochemistry, alteration — so you can see why it ranks where it does, and take it straight to your board or JV partner.
Signal from depth, not just the surface
We fuse geophysics, geochemistry and mapped geology with satellite data — so targets track where the deposit is, not only what happens to be exposed at surface.
An honest confidence flag
Every target carries an analogue-support signal — strong match, partial, or unproven ground — so you know which leads to back, and where a few field samples would sharpen the model.
Pick an area and a commodity. Here's what you get back.
In your workspace — in formats your GIS and your geologist can both open.
A map of ranked drill targets
Every promising spot on your ground, scored and ranked from best to worst.
The evidence behind every target
Host rock, geophysics, geochemistry, alteration — laid out so your team can check our work.
A trust flag on every target
Solid evidence, thin evidence, or a guess — so you know which targets to back and which to skip.
A drilling shortlist in priority order
Your best targets first, balancing how promising each one is against how sure we are.
A report you can hand over
Plain English for your board, your JV partner and your geologist — every line backed by the evidence.
Your data stays yours.
Anything you upload is used for your job only — never put into anyone else's model.
Two ways we model your ground.
Run a ready-made national model on any area today — or have a custom model trained on your own ground and data. We'll tell you which fits your project.
National models
Our production models, one per mineral per country — trained on the full national geoscience and ready to run on any area you draw. Every target comes with a plain reasoning chain and a confidence flag.
When: Best for scanning new ground fast, and when you need every target explained to a board or JV partner.
Learn moreCustom models
A model trained for your project — the national evidence stack plus your own samples and drilling. It sharpens the picture exactly where you work.
When: Best when you have local data to train on and accuracy on your ground is what matters most.
Learn moreBoth run on the same national dataset, both are tested the same honest way, and both flag the ground where they're guessing. Not sure which you need? That's the first thing we work out with you.
The model gets sharper on your ground.
Rank the ground, show how confident each call is, then target the data that would sharpen it most — and repeat.
Rank the ground
Every target is scored across the fused evidence — geology, geophysics, geochemistry and satellite — into ranked drill targets.
Flag the confidence
Each target carries a calibrated probability and an honest analogue-support signal — where the evidence is strong, and where the model is still guessing.
Sharpen with new data
Target a few field samples — or upload your own assays — where evidence is thin, and the model recalibrates to your ground.
What we model, where.
Certified national coverage for 51 minerals in three countries — 45 in Australia, 51 in the United States and 37 in Canada — plus 44 minerals you can run anywhere else in the world. A mineral is listed for a country only where a trained national model has passed our release gates and is live in the product.
| Mineral · 51 in the catalogue | Australia | United States | Canada | Everywhere else |
|---|---|---|---|---|
| Antimony | ✓ | ✓ | ✓ | ✓ |
| Barium | ✓ | ✓ | ✓ | ✓ |
| Beryllium | ✓ | ✓ | ✓ | ✓ |
| Bismuth | ✓ | ✓ | ✓ | ✓ |
| Boron | ✓ | ✓ | — | ✓ |
| Caesium | ✓ | ✓ | ✓ | ✓ |
| Chromium | ✓ | ✓ | ✓ | ✓ |
| Cobalt | ✓ | ✓ | ✓ | ✓ |
| Copper (porphyry systems) | ✓ | ✓ | ✓ | ✓Copper |
| Dysprosium | ✓ | ✓ | — | ✓ |
| Europium | — | ✓ | — | ✓ |
| Fluorine | — | ✓ | ✓ | — |
| Gadolinium | ✓ | ✓ | — | ✓ |
| Gallium | ✓ | ✓ | ✓ | ✓ |
| Germanium | ✓ | ✓ | ✓ | ✓ |
| Gold | ✓ | ✓ | ✓ | ✓ |
| Hafnium | ✓ | ✓ | ✓ | ✓ |
| Indium | ✓ | ✓ | — | ✓ |
| Lanthanum | ✓ | ✓ | — | ✓ |
| Lead–zinc–silver | ✓ | ✓ | ✓ | ✓Zinc |
| Lithium | ✓ | ✓ | ✓ | ✓ |
| Manganese | — | ✓ | ✓ | — |
| Molybdenum | ✓ | ✓ | ✓ | — |
| Neodymium | ✓ | ✓ | — | ✓ |
| Nickel–cobalt | ✓ | ✓ | ✓ | ✓Nickel |
| Niobium | ✓ | ✓ | ✓ | ✓ |
| Palladium | ✓ | ✓ | ✓ | ✓ |
| Phosphate | — | ✓ | ✓ | — |
| Platinum | ✓ | ✓ | ✓ | ✓ |
| Praseodymium | ✓ | ✓ | — | ✓ |
| Rare earth elements | ✓ | ✓ | ✓ | ✓ |
| Rhenium | ✓ | ✓ | — | — |
| Rubidium | ✓ | ✓ | ✓ | ✓ |
| Samarium | ✓ | ✓ | — | ✓ |
| Scandium | ✓ | ✓ | ✓ | ✓ |
| Selenium | ✓ | ✓ | — | ✓ |
| Silver | ✓ | ✓ | ✓ | ✓ |
| Strontium | ✓ | ✓ | ✓ | ✓ |
| Tantalum | ✓ | ✓ | — | ✓ |
| Tellurium | — | ✓ | ✓ | ✓ |
| Terbium | ✓ | ✓ | — | ✓ |
| Thorium | ✓ | ✓ | ✓ | ✓ |
| Tin | ✓ | ✓ | ✓ | ✓ |
| Tin–tungsten | ✓ | ✓ | ✓ | — |
| Titanium | — | ✓ | ✓ | — |
| Tungsten | ✓ | ✓ | ✓ | ✓ |
| Uranium | ✓ | ✓ | ✓ | ✓ |
| Vanadium | ✓ | ✓ | ✓ | ✓ |
| Ytterbium | ✓ | ✓ | — | ✓ |
| Yttrium | ✓ | ✓ | ✓ | ✓ |
| Zirconium | ✓ | ✓ | ✓ | ✓ |
Model skill is measured nationally on held-out known deposits — it is model-level skill, never a specific site's measured accuracy, and never a discovery or resource claim. A mineral is offered in a country only where its national model passed our release gates.
Now you can run ground anywhere on Earth.
Beyond our three certified countries, MineDSS reads your ground from the same trained evidence — 44 minerals, anywhere in the world, priced the same way. Draw a licence block in any country and see the prospectivity, the geology driving it, and how far to trust it — before anyone flies.
What it reads, where it's strongest, and how it differs from a certified national model.
Every target tells you how strong the evidence is.
Every target tells you how closely it matches ground we already know hosts deposits — so you can see your strongest leads, and where a few field samples would sharpen the model.
Analogue-matched
Looks like ground we already know hosts deposits. Your strongest leads.
Partial analogue
Part of the signature matches. A solid lead worth weighing.
Ground-truth priority
We haven't seen ground quite like this yet. A few cheap field samples and we calibrate the model to it.
No bluffing, no scary red maps — just an honest read on where your strongest leads are. Where the model wants more, a custom model on your own samples sharpens it.
The test that actually means something.
A prospectivity model only means something if it's tested on ground it never saw in training. Numbers measured on the same data a model learned from always look impressive — and tell you almost nothing about your tenement.
We do it the hard way. We hide the known deposits, build the model without them, then check whether it finds them anyway — with the test ground kept spatially separate, so the model can't cheat by memorising nearby points. That's the only test that tells you how it'll do on your ground.
And it has teeth: a model that doesn't pass isn't offered. For any mineral, in any country — if the national model didn't clear our release gates, you won't find it in the catalogue.
Tested on hidden deposits
We hide known deposits, rebuild the model without them, and test whether it still finds them — with test ground kept spatially separate so the model can't memorise nearby points.
Gated before release
Every model passes technical and performance gates before it is served. A model that doesn't pass isn't offered — for any mineral, in any country.
Honest on your ground
Every run carries a per-target analogue-support read — strong match, partial, or unproven ground — so the map tells you how far to trust it, target by target.
Simple, per-area pricing.
Pay for the ground you scan — no subscription, no seat licences, no lock-in. The exact price is shown before you pay, every time.
One mineral
Scan any area for a single mineral at $5 per square kilometre.
Add minerals to the same scan
One scan can score many minerals over the same ground — and each added mineral costs less than the one before. Six minerals run about $10.38/km², not $30.
Minimum billed area
Smaller areas run fine — they're billed at the 25 km² minimum. Priced in your country's currency: AUD in Australia, CAD in Canada, USD in the US and elsewhere.
Example: a 250 km² tenement scanned for gold costs $1,250. The same area scanned for six minerals at once — gold, copper, silver and three more — costs about $2,595, not $7,500. Every scan returns the full map, the evidence behind it, the confidence flags and the report — with ranked drilling targets in Australia, the USA and Canada.
Built on the public geoscience of three countries.
We've pulled together and cleaned the public geoscience across Australia, the USA and Canada — so your ground is scored against the full picture, not a thin local slice. Paired with satellite spectral, radar and terrain data available worldwide, that same foundation is what lets our global model read ground in any country on Earth.
soil & rock samples
mapped rock units
known occurrences & mines
mineral fingerprints
national geophysics layers
countries live · expanding
All public data — Geoscience Australia, the United States Geological Survey, the Geological Survey of Canada, and the state and provincial geological surveys. No proprietary lock-in. Kept up to date.
More on the data: remote sensing & Earth observation · how we fuse the layers · training data & validation
How it works
Self-serve, no lock-in. Draw an area, run it, explore the targets — right in your browser.
Draw your area
Draw any area of interest and pick your minerals — in our certified countries, or anywhere else in the world.
The engine runs it
MineDSS scores your ground against the full national dataset in the cloud, target by target.
Explore your results
The prospectivity map, the evidence behind it and the confidence flags — plus ranked drill targets in Australia, the USA and Canada.
Bring your own data
Upload your assays or drillholes and the model sharpens for your ground — used for your job only.
Straight talk: what this is, and what it isn't
- • MineDSS ranks ground and shows the evidence. It's there to focus your exploration — not a statement of fact.
- • It is not a resource or reserve estimate (JORC / NI 43-101), and not drilling or investment advice.
- • We make no discovery claims. We rank ground and show our reasoning; the drill confirms.
- • Where we're guessing, we say so — we don't hide it.
Better targets. Evidence you can check.
Drilling in Australia, the USA or Canada? Draw your ground and see the ranked targets and the evidence. Exploring anywhere else on Earth? Run it with our global model — and judge the output, not the pitch.
Sign up freeFree to sign up.