Built by a caver,
for cavers

CaveFinder was built by Zach Englebert: a caver, National Speleological Society (NSS) Director, and former cave inventory technician who spent years ridgewalking before wondering if terrain data could make every trip more productive.

In 2019, Zach served with the Montana Conservation Corps to search for caves in the Pryor Mountains of Montana and Wyoming for the US Forest Service and Bureau of Land Management. The crew documented over 150 karst features that summer, and the work was published in NSS News. But the process was the same one cavers have used for decades: pick a ridge, walk it, hope for the best.

After that summer (and years of professional karst survey work in New Mexico using drones, ground-penetrating radar, and GIS) the question became obvious: high-resolution LiDAR data exists for most of the country at 1 m resolution. It sees every dip, depression, and terrain feature invisible from the surface. Why wasn't anyone using it to find caves systematically?

CaveFinder is the answer to that question. It automates the terrain reading that experienced cavers do intuitively, but does it across an entire landscape in minutes.

Traditional cave hunting means driving to an area, picking a ridge, walking it, and hoping. Hours of bushwhacking through poison ivy to find a tree fall. Or nothing. Meanwhile, high-resolution terrain data sits unused because there's no easy way to analyze it.

The USGS has been systematically collecting LiDAR data across the country for years, creating incredibly detailed elevation maps that can reveal features invisible to the naked eye. But accessing, processing, and interpreting that data requires GIS expertise and significant computing resources that most cavers don't have.

CaveFinder automates what experienced cavers do intuitively (reading terrain for signs of subsurface voids) but does it systematically across an entire landscape in minutes.

Measured on 2,691 known US caves in states the model never trained on, the real cave is among the ten top-ranked spots 35% of the time, the top 25 47% of the time and the top 50 60% of the time (within 50 m). Full breakdown on the methodology page. Draw a box on the map, and the system downloads elevation data, runs the analysis, and returns ranked candidates, plus terrain visualizations.

Not every candidate is a cave, and we're upfront about that. But by narrowing thousands of acres down to a handful of the strongest leads, you can make every field trip count.

CaveFinder is a product of Buzzy LLC, based in Alabama. It wasn't built by a startup team that Googled “what is a cave.” It was built by someone who has mapped caves in southeastern New Mexico, fabricated bat-protective gates for endangered species in NM, AL, WV, and NY, and walked more ridges than he can count.

Every feature was designed around real field workflows: from the Ridgewalk Planner that generates printable route PDFs, to field status tracking (Cave found, Not a cave, Needs another visit) across trips. If it's in the tool, it's because it was needed in the field.

Conservation comes first. Cave ecosystems are fragile and irreplaceable, so we design our tools to support responsible exploration; read more on our conservation page. Privacy is the default: your analyses, candidates and field data stay private to your account, and our privacy policy is short and written in plain language.

We are honest about limitations. We show our detection rates, not just the wins, and not every candidate is a cave. And everything is built for the field, around real caving workflows: printable maps, GPS-ready tracks and results that work on a phone.