Title examination · private beta
Faster title examination. The examiner still makes the call.
We turn a title search package into a structured exam file. Every observation cites the document and page it came from. The examiner signs off.
We don't replace the title examiner. We give them their afternoon back.
How it works
Search package in. Exam file out.
Upload the search package. Deeds, mortgages, releases, liens, lis pendens, tax bills, all in one PDF. The engine pulls each instrument out, reads it, and produces an exam file organized the way examiners already read them. Every observation cites the document and the page.
Why us
What's actually different about this.
The hard part of title examination isn't reading the deed. It's knowing which facts to put in front of the examiner, and how to hedge. Most software gives you the data. We give you the read.
Practice, encoded
The patterns examiners look for, written down as rules the engine applies on every package. Not a generic chatbot dispensing legal opinions.
Examiner makes the call
Every callout reads "here is what we found, please confirm." Nothing reads "this title is clean." That's the examiner's job, not ours.
Per-instrument extraction
A bundled 100-page PDF becomes one card per recorded instrument. Deeds, mortgages, releases, assignments, liens, lis pendens, tax bills, each parsed and cross-referenced.
Private by default
Documents stay on the agency's hardware; cloud routing is opt-in per tenant and fails closed to local. SSN and date of birth are detected by presence only, never captured, and the order's transaction details never leave your server. The cloud destination, if you opt in, is SOC 2 Type 2 certified.
A starting point that deepens
Onboarding loads the rule sets for your states first. The knowledge base is a starting point, not a finished product: curation deepens it over time, and a loop that feeds examiner corrections back into the rules is on the roadmap.
Transparent per-order pricing
Each order returns a per-order cost breakdown (OCR pages, model tokens, render time), so you can see what drove the cost. Pricing adapts to volume and product mix; we discuss that with you, not on the marketing page.
Familiar layout
Four sections. Deeds. Mortgages. Judgments and Liens. Taxes. Each parent shows dependent documents nested below it. The format examiners have read for decades.
Foreclosure and divestiture checks
Completed foreclosures get surfaced. IRS §7425(b) 120-day redemption windows get surfaced. MERS-to-trustee assignment chains get surfaced. Dissolved-entity grantors get surfaced.
Honest about OCR
When a page can't be read cleanly, the engine flags it for examiner review with the exact page numbers. No confident-looking extractions from watermarked scans.
Typed extraction
Every field the engine emits is type-checked against a schema. Made-up fields get rejected. Missing required fields go to needs-review. Drift can't hide.
Get on the early-access list.
We'll reach out when there's something worth showing you.