Extracting data from allonges to notes
When a mortgage note is sold, the seller often endorses it on an allonge, a sheet attached to the note on its own letterhead or a plain form. Document custodians and foreclosure counsel review one or more in every transferred loan file, scanned with the note into the collateral file.
A custodian needs the note the allonge belongs to, identified by borrower, amount and property, plus the payee and the endorser. Collateral review matches each allonge to its note and confirms the endorsements run unbroken to the current holder.
What makes allonges to notes hard
Tied to its note by details
Fannie Mae requires an allonge to identify its note by borrower, note date, amount and property address, and each has to match the note.
Blank or specific endorsement
A specific endorsement names its payee, while a blank one is a bare signature that makes the note payable to bearer.
A chain across several sheets
A note sold more than once carries a stack of allonges, and the endorsements have to run in order from the originator to the holder.
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 allonge to note, end to end
About this document: An allonge to a VA-guaranteed mortgage note, endorsing it from PennyMac Corp. to PennyMac Loan Services, LLC, filed with a foreclosure in the Multnomah County Circuit Court. One scanned portrait page: six labelled lines, a signature crossing the signer's line and a hand-filled exhibit stamp. Page 1 is featured. The borrower and property were replaced with stand-ins, and the loan number was blacked out before filing.
Processing objective: The objective is to capture the transfer: which note the allonge belongs to, and who endorsed it to whom. The note is identified by its property and its principal, and the endorsement passes the note without recourse from one PennyMac entity to another.
Demonstration fields: Property address, Principal balance, Pay to the order of, Endorser (without recourse)
The document
1 page
Page 1shown below
Where each value came from
property_address
{
"value": "100 SAMPLE AVE , PORTLAND, OR 97000",
"page": 1,
"range": {
"end": 98,
"start": 44
},
"box": {
"xmin": 0.16014,
"ymin": 0.28361,
"xmax": 0.58415,
"ymax": 0.29844
}
}principal_balance
{
"value": 360000,
"page": 1,
"range": {
"end": 127,
"start": 100
},
"box": {
"xmin": 0.16076,
"ymin": 0.34161,
"xmax": 0.39035,
"ymax": 0.35685
}
}pay_to_the_order_of
{
"value": "PennyMac Loan Services, LLC",
"page": 1,
"range": {
"end": 215,
"start": 167
},
"box": {
"xmin": 0.16132,
"ymin": 0.48546,
"xmax": 0.51635,
"ymax": 0.50212
}
}endorser
{
"value": "PennyMac Corp.",
"page": 1,
"range": {
"end": 250,
"start": 218
},
"box": {
"xmin": 0.16191,
"ymin": 0.6032,
"xmax": 0.41891,
"ymax": 0.61837
}
}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.
