The structured intelligence layer for the mining industry.

Technical reports, financial filings, press releases, corporate registries, regulatory disclosures — structured into a single queryable database and growing every day. Ask questions about the mining industry that have never been answerable before.

$ minesignal query "systematic bias across QPs in metallurgical recovery assumptions for Carlin-type gold deposits in the western US"
The Database
100
Relational tables
And counting
2,600+
Structured parameters
And counting
Continuous
Data ingestion
New sources added daily
Data Sources
Every public disclosure. One structured database.
Technical Reports NI 43-101, JORC, and SEC/Edgar filings
Press Releases Classified and structured at scale
Financial Filings Corporate financials and transactions
Corporate Registries Officers, directors, and corporate networks
Regulatory Disclosures Permits, environmental, compliance
Scientific Literature Economic geology publications and research
What the database unlocks
Ask what was never answerable before.
01 Where do QP metallurgical recovery assumptions systematically diverge from actual plant performance — by deposit type, region, and author?
02 Which prospect generators are consistently securing the most favorable JV deal terms, and what do their earn-in structures look like?
03 What is the actual exploration cost per ounce discovered across western US gold projects over the past decade?
04 How do geotechnical design assumptions in pre-feasibility studies compare to as-built conditions across open pit operations?
05 Which teams keep working together across projects — and what do their projects actually produce?
The Alpha Signal
A 4% recovery miss cascades through everything.
Mining investors have never had a systematic way to audit the assumptions baked into feasibility studies — until now.
Case Study — Porphyry Copper
A feasibility study assumes 90% Cu recovery. The plant delivers 86%.
That 4% gap doesn't stay in the mill. It cascades through every cost structure the company has: less copper per ton of ore processed means less revenue. The same tonnage moves through the crusher, the SAG mill, the flotation circuit — same energy, same reagents, same labor — for fewer pounds of payable metal. Waste dumps grow faster. Strip ratios get worse. The mine plan that penciled at $2.50/lb Cu now needs $2.85 to break even.
-4.4%
Revenue per ton
+14%
Unit milling cost
+$0.35
Breakeven shift /lb
-31%
After-tax NPV impact
This pattern is not hypothetical. It repeats across porphyry copper projects globally. The question is whether the QP who signed the report has a track record of optimistic recovery assumptions — and whether the metallurgical testwork actually supports the number.
Now imagine tracking this across every deposit type, every QP, every study.
By structuring the assumptions inside thousands of technical reports — recovery rates, strip ratios, head grades, processing costs, capex estimates — and comparing them against actual production outcomes, MineSignal can identify systematic bias by author, by firm, by deposit type, and by region. These are alpha signals that exist in the public record but have never been queryable. We're making them queryable.
How We Build
AI is the tool. Not the product.
Domain-First Architecture Structured by geologists
The schema is designed by people who have written technical reports, not by people who have read about them. Every table, every column, every validation rule reflects how the data is actually used in practice.
Compounding Data Advantage Grows with every engagement
Every report our teams produce feeds structured data back into the database. The more we work, the more complete and valuable the dataset becomes. This advantage compounds and cannot be replicated overnight.
AI as Tailwind Accelerant, not substitute
We use AI to extract, classify, and structure data at a scale that would otherwise require hundreds of analysts. But every model is validated against domain expertise. The intelligence comes from the data and the people — AI makes it possible to build at this pace.
Products
Live
Verify
Automated QA/QC for drillhole assay data. Every element, every interval, every check — in seconds. Free.
Building
Intercept
Compositing, geochemistry flagging, and certificate linking. The paid tier.
Founding Team
Mining experts first. Software engineers second.
CEO
Brad Cantor
M.Sc. Economic Geology, University of Nevada — Center for Research in Economic Geology, an industry-funded graduate program. Background in exploration geology and drill program operations with a career-long focus on the junior mining space. 12+ years building and operating a manufacturing business at scale.
CTO
Taylor Kilian
PhD Geophysics, Yale University. Former KoBold Metals, where he worked at the intersection of machine learning and geoscience data at the company that proved the proprietary-database approach works at scale.
2,000–4,000
Technical reports filed globally per year
$50K–$5M
Cost to produce a single report
~$1.4B
Mining consultancy market by 2031
How to Think About Us
Bloomberg built the financial data terminal. Verisk built the insurance risk database. LexisNexis built the legal research layer. The mining industry — a trillion-dollar global market — has never had its equivalent. That's what we're building.