Extracting data from bailee letters
A warehouse lender sends a bailee letter with every mortgage note it ships to an investor. Each warehouse bank writes its own form. The investor's document custodian receives one with each loan it buys, as paper in the note shipment and as copies by fax or email.
An investor needs the loan number, the warehouse lender, the purchase price and the funding deadline. A collateral tracking system matches the letter to the note on file, and the funding desk wires the price to the warehouse lender before the window closes.
What makes bailee letters hard
Parties identified in legal prose
A bailee letter names its parties by role inside legal paragraphs: the warehouse lender, the originator and the investor acting as bailee.
A form per warehouse lender
Fannie Mae's Selling Guide defines a bailee letter by its function, so each warehouse lender drafts its own wording and layout.
Two amounts, one wire
A bailee letter can print the loan amount on its schedule and the purchase price to wire in its body, as two different figures.
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 bailee letter, end to end
About this document: A synthetic bailee letter that a document AI vendor publishes as its sample, shown unredacted because the parties and loan are invented. Four portrait pages: the letter across pages 1 and 2, a signed acknowledgment on page 3, and a one-loan Schedule A on page 4. Page 1 is featured.
Processing objective: The objective is to capture the deal the letter puts on hold: the loan, the parties on each side of it and the deadline for payment. Page 1 names the warehouse lender inside a legal paragraph, and spells out the funding window in words.
Demonstration fields: Loan number, Letter date, Investor (payee), Warehouse lender, Funding window (business days)
The document
4 of 4 pages
Page 1shown below
Page 2
Page 3
Page 4
Where each value came from
loan_number
{
"value": "123456789",
"page": 1,
"range": {
"end": 36,
"start": 16
},
"box": {
"xmin": 0.1021,
"ymin": 0.13469,
"xmax": 0.27765,
"ymax": 0.14957
}
}letter_date
Printed on the page as Date: 12/12/2024, normalized on extraction.
{
"value": "2024-12-12",
"page": 1,
"range": {
"end": 76,
"start": 60
},
"box": {
"xmin": 0.10173,
"ymin": 0.20879,
"xmax": 0.24223,
"ymax": 0.2232
}
}payee.name
{
"value": "Fake Home Mortgage",
"page": 1,
"range": {
"end": 111,
"start": 78
},
"box": {
"xmin": 0.10172,
"ymin": 0.24514,
"xmax": 0.40645,
"ymax": 0.26199
}
}warehouse_lender_name
{
"value": "ABC Sample Inc.",
"page": 1,
"range": {
"end": 570,
"start": 477
},
"box": {
"xmin": 0.0996,
"ymin": 0.41017,
"xmax": 0.89254,
"ymax": 0.42761
}
}funding_expiration_days
{
"value": 10,
"page": 1,
"range": {
"end": 1891,
"start": 1796
},
"box": {
"xmin": 0.09952,
"ymin": 0.68459,
"xmax": 0.85494,
"ymax": 0.70277
}
}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.
