Extracting data from student loan statements
Loan servicers and government agencies issue student loan statements in their own layouts, from US federal servicers to New Zealand's Inland Revenue. Mortgage lenders request one when the credit report needs support, such as a payment in deferment or reported as zero, usually as a PDF from the servicer's portal.
Underwriting needs the current balance, the monthly payment and the repayment status. The loan origination system counts the payment in the debt-to-income ratio, or a percentage of the balance where the statement shows a payment of zero.
What makes student loan statements hard
Several loans on one account
A US federal servicer often lists each disbursed loan separately under one account, each with its own balance and interest rate.
Status changes the payment
A loan in deferment or forbearance can show a payment of zero, with the date that status ends printed in a separate section.
Servicers in many countries
A statement from a US federal servicer and one from a national tax agency differ in layout, date format and the identifier they carry.
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 student loan statement, end to end
About this document: A New Zealand student loan statement from Inland Revenue, the government's tax agency. Three portrait pages: a balance callout, then tables of increases and decreases to the loan, and notes on repayment. Page 1 is featured. Every name, address and identifier was replaced with stand-ins before parsing.
Processing objective: The objective is to see how the balance was reached. Course fees and living costs add to the loan, interest and salary deductions come off it, and the current balance is what remains. Each value grounds to its own line, sign intact.
Demonstration fields: Compulsory course fees, Living costs, Interest, Salary and wage deductions, Current loan balance
The document
3 of 3 pages
Page 1shown below
Page 2
Page 3
Where each value came from
increases[0].amount
{
"value": 13509.6,
"page": 1,
"range": {
"end": 920,
"start": 910
},
"box": {
"xmin": 0.32415,
"ymin": 0.47834,
"xmax": 0.90942,
"ymax": 0.49962
}
}increases[2].amount
{
"value": 2958.48,
"page": 1,
"range": {
"end": 1026,
"start": 1017
},
"box": {
"xmin": 0.32404,
"ymin": 0.52127,
"xmax": 0.90938,
"ymax": 0.54419
}
}decreases[1].amount
{
"value": -589.12,
"page": 1,
"range": {
"end": 1345,
"start": 1337
},
"box": {
"xmin": 0.32367,
"ymin": 0.6837,
"xmax": 0.90921,
"ymax": 0.70581
}
}decreases[2].amount
{
"value": -2046.12,
"page": 1,
"range": {
"end": 1409,
"start": 1399
},
"box": {
"xmin": 0.32361,
"ymin": 0.70573,
"xmax": 0.90918,
"ymax": 0.73197
}
}current_loan_balance
{
"value": 26452.96,
"page": 1,
"range": {
"end": 1528,
"start": 1468
},
"box": {
"xmin": 0.08882,
"ymin": 0.76511,
"xmax": 0.9105,
"ymax": 0.78641
}
}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.
