otherwatersflow

That's fair, I'd like to see which ones. Listings update once a week (Mondays), so some will have sold since; each page shows when it was checked. The price is the listing's asking price, not an estimate. A gap like $59k vs $470k usually means one site just land and the other has a house. Or they matched different parcels. If you can paste a few of links I'll trace them.

Thanks for the example, and congratulations on the land. That sounds like exactly the kind of place this is for. That listing isn't in my data at all, so it's a coverage gap rather than a measurement miss: either it wasn't in this week's listings feed or it was listed after Monday's update. I'll trace it this weekend.

That's exactly it, thanks for explaining it better than I did. I ran into this early with Vermont ski-resort condos: every unit "had" the acreage of the whole resort. The lot-size field just doesn't fit condos well. The site already hides acreage on listings typed as condos and flags units that resolve to an association's common parcel, but some condos come through typed as townhouses or single-family, and those slip past. Checking the claimed acreage against the actual parcel size is the fix I'm working on.

Thank you, that's great to hear. Honestly, the number of properties next to BLM was one of the things that surprised me most when I started building this, there's ~2,700 right now. On the map you can set the land type to BLM to see just those. And since you're always looking: any search can send you a weekly email when new matches turn up (starts Monday). Curious what else you check for on BLM-adjacent land?

Good one, thanks. I think OpenStreetMap has decent bike-trail coverage, so "distance to the nearest trail" is quite doable. It's on the list.

Not yet. The public-land source (USGS PAD-US) does include county/city land, but most of it is small parks, and I set a 50-acre minimum. County forests and large preserves are a good idea though, I'll look at adding those.

Not that one. Parcels come from about 20 states' own statewide GIS services, and from Realie (licensed) for the rest. I'd want to check the license on that Kaggle dataset before using it, since county parcel data comes with all sorts of terms.

Same here, I found my land bordering state game lands in the Poconos for exactly that reason. I'm sorry to hear about the watershed woods; that's the risk with land that isn't actually protected. Fracking leases and mineral rights are a great point too. They're hard to see from public data, but worth thinking about.

Thank you, that's exactly the kind of report I need. Acreage comes from the listing data, which is messy: lot sizes arrive in square feet, or as the whole condo association's land, and my checks catch a lot of it but clearly not all. Could you share the links (here or contact@unfenced.land)? I'll trace both and fix the rule that let them through, not just those two listings.

No, they're in. About 470 properties border or sit near NPS land (parks, monuments, recreation areas, lakeshores), e.g. a few along the edge of the Smokies. You're right that the headline copy only mentions national forest, BLM and state land, so that's misleading; I need to update the wording. The map's land-type filter has a "National Park Service" option if you want to see just those.

Very true. Bordering a big block of public land doesn't mean you can walk through it easily, and the acreage number says nothing about brush or deadfall. That's part of why the site only reports what it can measure (the shared edge, the size, whether a mapped road reaches the parcel) and doesn't promise anything about getting around once you're out there. Trail data would be a nice addition someday.

That matches my experience next to state game lands in PA. There's a few hunters in season who know where the lines, especially with OnX, Mostly quiet the rest of the year.

That's the flip side of this: private land controlling the way into public land. It's a real problem in a lot of places, and an interesting thing to map someday.

American Prairie in Montana is an example going the other way: a nonprofit buying private land that surround public land and opening them to the public. It's somewhat contentious locally, but it shows how much the private parcels around public land shape who can actually use it.

"Nobody builds up the road" is the reason I hear most often from buyers, and it's personal for me too. My aunt and uncle bought a home in rural Ohio 35 years ago, on about a few acres, in the middle of nowhere. Over the years they've watched the farmland around them get sold off and housing developments creep closer.

Land that borders public land doesn't have that problem on at least one side. That's also why the site shows which agency manages the land next door, since that affects how likely it is to stay as it is.

Thanks, this is useful feedback.

On people crossing over: it depends a great deal on the land. Next to state game lands you'll see hunters in season; next to a big national forest you may rarely see anyone. It's very dependent on individual perspective and location.

Bordering vs adjacent: bordering means the parcel lines actually touch and there's a measured shared edge; adjacent means no shared edge but within about 250 ft (usually a road or a strip of other land in between). You're right that the state pages don't make that clear, and feet is the better unit for a US site. Tho most US geospatial data is published in metric lol.

And fair point on the copy. I run a small design studio and built this in my free time, I leaned on AI for a lot of the site text. The "Why this exists" section on How it works is my own story, but the rest deserves a pass in my own voice.

Mostly Claude Code. I have a background in 3D Design and Geospatial Mapping, the data sources and the decisions about what counts as "bordering" were mine

Great list, and it's pretty much what serious land buyers end up checking by hand. A few are already there: elevation, whether a mapped road reaches the parcel, FEMA flood zone and wildfire hazard. Several more look doable from public data: slope (especially near the road) from the same USGS elevation model, days above 85F/below 30F from NOAA climate normals, and distance to hospitals, groceries, airports and shooting ranges from public facility datasets. Power availability is the hard one, since distribution lines mostly aren't public.

If you had to pick the three that rule a parcel in or out fastest, which would they be? That'd help me decide what to build next.

Good one. Wetlands are a gap right now. The closest thing the site shows is the FEMA flood zone for each parcel, which catches some wet ground but not a hidden pond or swamp like yours. I believe US Fish & Wildlife Service publishes a National Wetlands Inventory for the whole country, so checking how much of each parcel is mapped wetland is very doable. Adding it to the list, thanks.

Thanks, and agreed, that's the main use case. The map view (unfenced.land/map) already has min/max price and acreage filters, plus property type, shared-boundary length, elevation and road access. The ZIP search and state lists can only sort by price and acreage right now, so adding the same filters there is a good call. It's going on the list.

Thanks, and this is a really useful perspective. The fence and dead-tree problem is exactly the kind of thing a map can't show.

My experience has been the opposite, though it's a different situation: I bought recreational land in Pennsylvania that borders state game lands, and I don't live there full time. What I wanted was the ability to step off my own ground into something wild and mostly untouched, and it's been a very positive experience.

I think you've put your finger on the real point: it's a trade-off, and it depends on what you want from the land. Living next to it full time with dogs is a very different thing from a weekend place you go to so you can disappear into the woods. "Go in with eyes open" is good advice either way.

For some buyers, yes, and that's a fair point. The site doesn't claim bordering public land is better, it just measures it, because listings say "borders state land" constantly and it's hard to check.

The people who've been most enthusiastic so far are hunters, homesteaders, off-grid folks, and people who want somewhere to walk out the back door and keep going, where the public land is the whole point. Others will see the same boundary as trespass risk, hunting pressure or a fence the state won't fix (see the comment above about a state park). Same fact, different buyer.

On liability, rules vary by state, so that's worth checking with a local attorney before buying, whatever the map says.

Thanks, that's really nice to hear. Happy to share:

- Listings: RentCast's API (licensed), pulled weekly.

- Parcel boundaries: state and county GIS services where they publish statewide parcels (about 20 states), and Realie (licensed) for the rest.

- Public land: USGS PAD-US 4.1, filtered to fee-owned, open-access land of 50+ acres.

- Elevation: USGS 3DEP point queries. Roads: Census TIGER. Flood: FEMA's National Flood Hazard Layer. Wildfire: Forest Service Wildfire Hazard Potential (2023). ZIP search: Census ZCTA centroids.

The free state parcel services were the biggest hunt: every state publishes differently, and a couple that looked open turned out to be test or internal-use-only endpoints.

It's RentCast (rentcast.io) listings API and parcel boundaries come from state/county GIS services and Realie (realie.ai). Fair question on the rules: I'm not a broker or MLS member, so NAR/MLS rules don't bind me directly; I license the listings through RentCast's API under terms that allow displaying them, and how they source the data is on their side. I also left agent phone numbers and emails out of the downloads on purpose so the site can't be used to build contact lists. If anyone from an MLS sees a problem with a specific listing, I'd genuinely like to hear about it: contact@unfenced.land

Hi HN, I built this. Listings that say "borders state land" can't easily be checked: you open the county's GIS viewer and line up parcel lines against public-land maps by hand. I did that a lot while looking for land of my own, then automated it.

How it works, once a week:

1. About 700,000 active U.S. listings come in from a licensed listings API.

2. Each listing near public land gets its county parcel boundary, from state and county GIS services where they publish them and from a licensed parcel provider elsewhere. With no boundary, the page says so and measures from the listing point, but never claims a shared edge.

3. Public land is USGS PAD-US 4.1, narrowed to land an agency actually owns (Category = Fee, so no conservation easements), that's open to the public, and that's at least 50 acres.

4. Parcel and public land are compared in a local metric projection (shapely/geopandas): within (an inholding), bordering (the boundaries touch, and the shared edge is measured in feet), or adjacent (within 75 m, about a road's width).

Each property also gets the size of the connected public land behind it, its elevation (USGS 3DEP), road access (Census TIGER roads), FEMA flood zone and Forest Service wildfire hazard. The output is about 16,000 static pages with Leaflet maps on Cloudflare Pages, rebuilt every Monday on a Mac mini. ZIP search looks the ZIP up in your browser from Census centroids, so nothing you type is sent anywhere.

Things that were harder than expected:

- Lot sizes sometimes arrive in square feet in the acres field. A Birmingham, AL listing came through at 7,000 acres. Sizes are now checked against price, property type and the parcel itself, and corrected or withheld with a note.

- Condo and subdivision listings often resolve to the association's common tract, so a single town lot can appear to share miles of boundary. Those are flagged as shared parcels.

- PAD-US stores some units as one multipart polygon. My first "connected public land" figure for one Montana property was 69.7 million acres, about three-quarters of the state. Units are now split into pieces and joined only where they actually touch.

- Road centerlines and lot lines come from different agencies and are often tens of meters apart, so "does a road touch this parcel" needs a tolerance, and frontage only counts where the boundary runs along the road.

It's free, with no account and no cookies. There are free weekly email alerts for any search, starting Monday. I'd especially like to hear from anyone who has bought rural land: what should it measure that it doesn't?