Extracting data from promissory notes
Every mortgage closes on a signed promissory note, usually a Fannie Mae or Freddie Mac uniform instrument such as Form 3200, with adjustable-rate and state variants. Document custodians certify one for every loan delivered into a pool or sold, reading a scan of the original in the collateral file.
Certification needs the principal, the interest rate, the monthly payment, the maturity date and the lender. The custodian matches them against the loan schedule, and a note whose principal or payment differs holds up certification of the pool.
What makes promissory notes hard
Terms written as sentences
A uniform note states the principal, rate and payment in first-person prose, such as "I will pay interest at a yearly rate of".
Small type on scanned copies
A note runs to three or more pages of small type, and the copy in a collateral file or court filing is usually a scan.
Near-identical form families
Fixed-rate, adjustable-rate and state-specific notes share most of their wording, and the adjustable versions add change dates and rate caps.
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 promissory note, end to end
About this document: A VA-guaranteed mortgage note on the Fannie Mae and Freddie Mac Multistate Fixed Rate Note, Form 3200, filed as an exhibit to an Oregon foreclosure judgment. Three scanned portrait pages of small type, initialled on each page and signed on the last. Page 1 is featured. The borrower's name, signature, initials and property were replaced with stand-ins, and the loan number was blacked out before filing.
Processing objective: The objective is to read the loan's terms: who lent the money, how much, at what rate and for what monthly payment. Each term is written into a sentence of the note's first-person prose, and each grounds to the sentence that states it.
Demonstration fields: Lender, Principal, Interest rate (%), Monthly payment
The document
3 of 3 pages
Page 1shown below
Page 2
Page 3
Where each value came from
lender
{
"value": "Mortgage Express, LLC",
"page": 1,
"range": {
"end": 532,
"start": 425
},
"box": {
"xmin": 0.22733,
"ymin": 0.34495,
"xmax": 0.78454,
"ymax": 0.35633
}
}principal_amount
{
"value": 360000,
"page": 1,
"range": {
"end": 424,
"start": 311
},
"box": {
"xmin": 0.24736,
"ymin": 0.33363,
"xmax": 0.81202,
"ymax": 0.34532
}
}interest_rate_percent
{
"value": 3.875,
"page": 1,
"range": {
"end": 971,
"start": 949
},
"box": {
"xmin": 0.22757,
"ymin": 0.44873,
"xmax": 0.34719,
"ymax": 0.45947
}
}monthly_payment
{
"value": 1692.85,
"page": 1,
"range": {
"end": 1972,
"start": 1913
},
"box": {
"xmin": 0.24809,
"ymin": 0.62261,
"xmax": 0.55735,
"ymax": 0.63481
}
}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.
