Extracting data from wire instructions
Title companies, law firms and loan servicers each print wire instructions on their own letterhead. A closing team receives one from every party it pays at a real estate closing, including the seller and each lender being paid off. They arrive as PDF attachments by email or through a closing portal.
Escrow needs the receiving bank, the routing number, the account number, the account name and the reference to quote. Those go into the bank's outgoing wire system once a callback to a number already on file confirms them.

What makes wire instructions hard
Forged copies look genuine
Before funds are wired, business email compromise sends a copy on the same letterhead with only the bank and account changed.
Each payee writes its own labels
The same field reads Bank Account on one payee's page and Beneficiary Account No. on the next, each in its own layout.
Lookalike numbers stacked together
Routing, account and reference numbers sit a line apart as similar strings of digits, and each one belongs in its own field.
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 wire instruction example, end to end
About this document: First Integrity Title Company's wire instructions for closing funds, one portrait page split from a preliminary title report. A letterhead and the terms for sending funds sit above the instruction block, which runs from the receiving bank to the property address. Page 1 is featured. The account, routing and order numbers and the property address were replaced with stand-ins.
Processing objective: The objective is to capture the full payment route, from the bank that receives the wire to the reference the sender quotes. Each value is grounded together with its printed label, which is the pairing a reviewer checks against a known-good copy before any money moves.
Demonstration fields: Receiving bank, ABA routing number, Account number, Account name, Reference order number
The document
1 page
Page 1shown below
Where each value came from

receiving_bank.name
{
"value": "First Western Trust Bank",
"page": 1,
"range": {
"end": 595,
"start": 562
},
"box": {
"xmin": 0.17283,
"ymin": 0.316,
"xmax": 0.60921,
"ymax": 0.33066
}
}aba_routing_number
{
"value": "796630226",
"page": 1,
"range": {
"end": 691,
"start": 665
},
"box": {
"xmin": 0.17214,
"ymin": 0.39,
"xmax": 0.49653,
"ymax": 0.40841
}
}account_number
{
"value": "6677087",
"page": 1,
"range": {
"end": 713,
"start": 692
},
"box": {
"xmin": 0.173,
"ymin": 0.41992,
"xmax": 0.47848,
"ymax": 0.43819
}
}account_name
{
"value": "First Integrity Title Company, a California Corporation - California IOLTA Account",
"page": 1,
"range": {
"end": 810,
"start": 714
},
"box": {
"xmin": 0.17294,
"ymin": 0.45001,
"xmax": 0.92103,
"ymax": 0.48177
}
}reference_number
{
"value": "FIN-78796415",
"page": 1,
"range": {
"end": 844,
"start": 811
},
"box": {
"xmin": 0.1724,
"ymin": 0.49521,
"xmax": 0.52099,
"ymax": 0.51225
}
}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.
