Questions we actually get

A web-based tool that analyzes LiDAR terrain data to detect potential cave entrances. You draw an area on the map, we analyze the terrain, and you get ranked candidate locations to investigate in the field.

Recreational cavers looking for new leads, grotto survey teams planning fieldwork, karst researchers conducting terrain inventories, geotechnical engineers assessing sinkhole risk, and anyone who wants to understand what the terrain is hiding.

No. The interface is designed to be approachable for anyone. But understanding how to safely visit candidates in the field requires proper training and equipment.

We validated CaveFinder blind against over 10,000 known cave entrance records across multiple karst types. Each analysis generates a ranked shortlist of candidates by confidence score. Measured at a 50 m match radius: 40.4% of known entrances land in the top 10 candidates, 59.5% in the top 25, and 73.7% in the top 50. That's orders of magnitude better than random chance. Not every candidate is a cave, and not every cave will be detected — but the ranking consistently puts real caves near the top of the list. Full breakdown on the methodology page.

We use publicly available LiDAR elevation data from the USGS 3D Elevation Program (3DEP) at 1-meter resolution for most of the US. For areas without LiDAR coverage, we fall back to 30-meter satellite DEMs (Copernicus, SRTM).

Yes, with reduced accuracy. 1 m LiDAR is primarily available in the US. International areas use 30 m satellite DEMs, which can detect larger features but miss smaller entrances.

CaveFinder runs 12 detection methods against public LiDAR elevation data and combines them through 14 calibrated scoring formulas. The full plain-English inventory of every component is on the methodology page.

That layer shows publicly documented caves from OpenStreetMap and Wikidata. It's separate from CaveFinder's terrain analysis — it's there for reference so you can see what's already known in your area.

No. Your analysis results are purged within 24 hours. Exports are generated on-the-fly and not stored. Your searches are private to your account.

Never. We don't share, sell, or aggregate user data. See our privacy policy for details.

No. CaveFinder analyzes terrain data only. The “Known Caves” layer shows only publicly documented caves from OpenStreetMap.

Yes. 3 analyses per week, top 10 candidates, no credit card required, no time limit.

Yes. No contracts, no cancellation fees. Your subscription ends at the end of the billing period.

All major credit cards, processed securely through Stripe.

If you're a research team, land agency, grotto, or consultant running systematic surveys at scale and the standard tiers don't fit, email zach@cavefinder.app and we'll work something out.

Yes. The web app works in any modern browser. For best results, use a tablet or desktop for the larger screen.

Not yet for the elevation source. But Pro users can upload their own GeoTIFF map layers (geology, aerials, karst inventories, etc.) as semi-transparent overlays aligned to the map — useful for cross-referencing results against data you already have. Pro caps: up to 8 overlays at a time, 500 MB per file, 1 GB total stored, 15 uploads per day. Overlays auto-expire after 24 hours — they're working data, not long-term storage.

All users can export to CSV (spreadsheet), KML (Google Earth), GPX (GPS devices), and GeoJSON. Free exports cover your top 10 candidates; Pro exports the full ranked list. The Ridgewalk Planner also generates a printable PDF with topo map (Pro only).

The default scoring is tuned for karst, where most caves form. Pro users can toggle a non-karst filtering mode that loosens the karst-specific weighting when you're working in talus, lava tubes, or other non-karst settings.