Our Story

Why we built MineralScout.

Mineral title research is manual, slow, and error-prone. It hasn't meaningfully changed in decades. We thought software could fix that.

The problem

Mineral acquisition is a pipeline business. Source leads, research title, identify owners, make offers, close. The math is simple: more parcels researched per month means more deals closed.

The bottleneck is always title research. A single parcel can take a landman hours, sometimes days. Search the county grantor-grantee index, trace every conveyance from patent to present, resolve fractional interests, flag probate gaps, compile a run sheet. It's meticulous, high-skill work. Almost entirely manual.

So brokerages end up constrained by research capacity, not deal flow. They have more leads than they can work. Every week, opportunities age out because nobody had time to run the title.

Why existing tools didn’t cut it

People have tried to bring technology to mineral research before. County clerk websites, data aggregators, various landman tools. We spent time with all of them. A few patterns stood out:

Rented data, limited control

Most platforms license their data from aggregators. They don't control the data model, update schedule, or coverage. If the aggregator doesn't have a county, neither do they. Errors propagate unchecked.

Partial coverage

Many tools start their records at whatever date the county went digital, often the 1980s or later. For mineral work, where severances go back to the early 1900s, that's a dangerous blind spot. A title search that misses a 1920 reservation is worse than no search at all. It gives you false confidence.

Search, not research

Most tools give you keyword search across a document index. That's useful, but it's not title research. Title research means walking the chain link by link, grantor to grantee, following every branch, resolving fractions, flagging gaps. A search box doesn't do that.

What we decided to build

We started with one conviction: if you want to automate mineral title research, you have to own the data. Not rent it. Not link to it. Own it, in your own database, parsed at the field level, indexed for graph traversal.

So that's what we did. We started with Weld County, the most active oil and gas county in Colorado. We ingested the complete county clerk grantor-grantee index, patent to present. Every recorded document. Every grantor, grantee, document type, recording date, and legal description, parsed and normalized and loaded into PostgreSQL.

Then we built on top of that. BFS chain traversal that walks the grantor-grantee graph link by link. Fractional interest math that tracks minerals as they're divided and recombined across transactions. Probate gap detection. Canonical party resolution that turns “Robert J. Smith,” “R.J. Smith,” and “Bob Smith” into one identity.

Then Adams County. Same process, same infrastructure. And we're continuing to expand.

The technical approach

A few technical decisions define what the product can do. Worth explaining because they're not obvious from the outside:

Own the data

County clerk records live in our PostgreSQL database, not behind a third-party API. We can build graph queries, run BFS traversals across the grantor-grantee network, and do things that require random access to the full dataset. You can't automate chain-of-title against a search API.

Parse at the field level

We don't store documents as blobs of text. Every record is broken into structured fields: grantor, grantee, document type, recording date, book/page or reception number, legal description. Legal descriptions get further parsed into normalized PLSS components. That structure is what makes automated chain traversal and fractional interest math work.

AI for judgment, not just search

Finding documents isn't the hard part. Making judgment calls is. Is this the same Robert Smith? Does this probate close the gap, or is there a missing heir? Is this reservation 1/2 of the minerals or 1/2 of the grantor's interest? We use AI for those decisions, not just to index text.

What this means if you’re a landman or mineral buyer

We built MineralScout because the people doing mineral acquisition deserve better infrastructure. Not prettier dashboards. Infrastructure that lets you run a title chain in minutes instead of days, that catches the 1920 reservation a partial-coverage tool would miss.

We're not done. We're expanding coverage, improving the AI, and building what our users tell us they need. If you work in mineral acquisition and want to see what this looks like in practice, we'd like to show you.