Terminus — SB 17 land ownership screening
Terminus was the Roman god of boundary stones. Moving one was a capital offence, because the boundary was the thing everything else rested on.
Engagement-priced, per screening.
The problem this exists for
Texas SB 17 restricts the acquisition of Texas real property by entities connected to designated countries. It carries criminal penalties, which changes what a screening tool is allowed to be. A tool that returns a confident answer built on incomplete records is not a convenience — it is a way for somebody to make a legal determination on evidence that will not hold.
So this page leads with what the data cannot tell you, before what it can.
What the corpus actually is
| Texas parcels | 13,124,564 |
| Counties covered | 257 |
| Parcels carrying an acquisition date | 1,521,024 (11.6%) |
| Parcels with an owner mailing state | 12,880,723 (98.1%) |
| Of those, mailing address outside Texas | 665,018 (5.2%) |
Three limits, stated plainly
Acquisition dates exist for 11.6% of parcels (1,521,024 of 13,124,564). SB 17 turns on when a holding was acquired, because the statute is not retroactive in the same way across all cases. For the other 88.4% (11,603,540 parcels) we cannot tell you when the property changed hands, and no amount of analysis manufactures a date that the county never recorded. A screening that quietly treats a missing date as "not recent" would be telling you something the record does not support.
An out-of-state mailing address is a signal, not a finding. 665,018 parcels have an owner mailing address outside Texas. The overwhelming majority are ordinary — a Texan who moved, a trust administered elsewhere, a lender, a company with an out-of-state registered agent. Treating that column as a proxy for foreign ownership would generate hundreds of thousands of false positives and bury the handful that matter. It is one input among several, weighted accordingly.
Thirty-seven acquisition dates were impossible, and we removed them rather than shipping them. They carried years including 2056, 6201, 7791 and 8201 — digit transposition during transcription, not a parsing fault. The original values are preserved so they can be corrected from source records; the date field is empty for those parcels rather than wrong. That is 37 rows of 1,521,024, and we mention it because a vendor who has never found an error like this in their own corpus has not looked.
What the screening produces
A structured record, not a verdict. Ownership as the county recorded it, the corporate or trust layer where one exists, the acquisition date where one exists, and the specific reason each parcel was surfaced — so a reader can disagree with any individual step.
Entity structure rather than name matching. Owner names are classified by structure — individual, corporate, trust, government — and corporate holdings are followed to the extent the public record allows. The classification is on the legal form of the entity, which is what the statute addresses.
No inference about who people are. No attribute of a person is inferred from their name, and none is used to select, score, rank, filter or target. That is a standing constraint across this platform, not a Terminus-specific policy, and it is audited in code rather than promised in a document. See why we are different.
What this is not
It is not a legal determination, and it is not an investigative service. This is research and analysis conducted against public records. Whether a particular holding violates SB 17 is a question for counsel, on a record that includes things no public dataset contains. We produce the analysis that counsel works from.
It is not a watchlist. The output is a screening record about parcels, scoped to a specific engagement.
How to test it
Bring a holding you have already researched and know the answer to. That is a better test than anything we could select for you, and it is how we would rather be evaluated — including on the parcels where our answer is "the county never recorded a date, and we will not guess."