Extracting data from deeds
Attorneys and title companies draft deeds in their own words, and each county recorder adds its own stamps and margins. Title examiners read every deed in a property's chain of title, for each sale and refinance their office closes. The deeds arrive as recorder downloads and scans of decades-old paper.
Title review needs the grantor, the grantee, the consideration, the transfer tax and the parcel number with its legal description. A title production system checks each conveyance against the chain, and the title commitment names the vested owner from the latest deed.
What makes deeds hard
The parties sit in prose
A deed names its grantor, grantee and price inside one granting clause of running text, each party followed by its entity description.
Legal descriptions on exhibits
The legal description, by lot and block or by metes and bounds, often sits on an exhibit page that the deed's face points to.
Decades of scanned paper
A chain of title reaches back decades, to photocopied deeds carrying recorder stamps and dates filled in by hand.
Parse, extract, verify
Parse
Parse turns the document into reading-order Markdown and a hierarchy of blocks (text, tables, figures), and grounds each block to its exact location on the page.
Extract
You define the fields in a JSON schema, and Extract returns their values in a consistent, structured format, each linked back to where it appears in the Markdown.
Verify
Each value links back to where it appears, down to the character range in the Markdown, and the page and bounding box in the source document, so every result can be verified.
One deed, end to end
About this document: A grant deed from the City of Los Angeles, kept in the City Clerk's public council file. It sells a surplus parcel in Pacific Palisades to a California corporation. Five scanned portrait pages: the deed, the Exhibit A legal description, approvals, the execution page and the notary acknowledgment. Page 1 is featured. Every party to the conveyance is a public body or a company.
Processing objective: The objective is to read the conveyance on the face of the deed, along with the parcel it covers and the county tax declared on it. A title examiner checks that the grantor matches the grantee of the prior deed in the chain. Each value grounds to its own line of the page.
Demonstration fields: Grantor, Grantee, Consideration, County transfer tax, APN
The document
4 of 5 pages
Page 1shown below
Page 2
Page 3
Page 4
Where each value came from
grantor.name
{
"value": "THE CITY OF LOS ANGELES",
"page": 1,
"range": {
"end": 486,
"start": 393
},
"box": {
"xmin": 0.04502,
"ymin": 0.35604,
"xmax": 0.90454,
"ymax": 0.37151
}
}grantee.name
{
"value": "WHITE PICKET FENCE, INC.",
"page": 1,
"range": {
"end": 660,
"start": 610
},
"box": {
"xmin": 0.27997,
"ymin": 0.40799,
"xmax": 0.75285,
"ymax": 0.42664
}
}consideration_amount
{
"value": 1100000,
"page": 1,
"range": {
"end": 582,
"start": 487
},
"box": {
"xmin": 0.046,
"ymin": 0.36967,
"xmax": 0.90334,
"ymax": 0.38533
}
}documentary_transfer_tax[0].amount
{
"value": 1210,
"page": 1,
"range": {
"end": 103,
"start": 59
},
"box": {
"xmin": 0.5131,
"ymin": 0.16052,
"xmax": 0.84337,
"ymax": 0.1761
}
}apn
{
"value": "4412-028-906",
"page": 1,
"range": {
"end": 391,
"start": 374
},
"box": {
"xmin": 0.57141,
"ymin": 0.30972,
"xmax": 0.71033,
"ymax": 0.32399
}
}How to try it yourself
Everything on this page is reproducible. The sample, its schema, and the Parse and Extract API outputs are published. You can repeat the same two calls from whichever surface you already build on.
- Claude Skilldescribe the job; the agent writes it
- CLIno code at all
- MCPexpose ADE docs as an agent tool
- REST APIany language
- Pythonlandingai-ade
- TypeScriptlandingai-ade
Documents into data
Accurate
The best accuracy on the difficult documents other tools get wrong, such as messy scans, complicated tables, checkboxes and handwriting.
Auditable
Every answer cites its sources. Values come back with an audit trail (page, bounding box, character range) so a human can review them.
Agentic
Applies multi-step reasoning, and plans before it acts. Handles documents it has never seen before and adapts to variations.
API-first
Composable APIs that drop into any tech stack and can be called from anywhere. Easy to build on and embed in your pipelines and products.
Affordable
Cost-competitive rates with complete price transparency. Spend controls before you run, and every job reports its exact cost.
Enterprise-grade
Built for production and scale from day one. Fully observable and monitored, running securely in our cloud or yours.
Built for high-volume, high-throughput production
- Zero data retention
- Priority rate limits
- Enhanced support
- VPC and on-prem deployment
Available on Team and Enterprise plans. VPC and on-prem deployment require an Enterprise plan.
