Extracting data from disability benefit letters
Disability benefit letters come from the Social Security Administration, the VA and private disability insurers, each in its own wording. Mortgage lenders and landlords collect one from every applicant who counts disability income, as a PDF downloaded from the payer's website or a scanned copy.
Underwriting needs the gross monthly benefit, any deductions, the net payment and the month the amounts took effect. The lender's income worksheet carries the gross benefit into the debt-to-income ratio, grossed up where the income is nontaxable.
What makes disability benefit letters hard
Every value inside a sentence
Amounts and dates sit in running prose, such as the premium deducted each month, each written into a sentence of explanation.
Current and past in parallel
Sections for current and past benefits repeat almost the same sentences, each with its own amounts and the months they apply to.
Each payer words it differently
A letter from the SSA, the VA or a private disability insurer names each amount in its own words.
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 disability benefit letter, end to end
About this document: A Benefit Verification Letter from the Social Security Administration for a Social Security Disability Insurance beneficiary. Two portrait pages of prose: current and past benefit amounts on page 1, then Medicare entitlement and contact details on page 2. Page 1 is featured. The recipient's name, address and identifiers were replaced with stand-ins, and the mailing barcodes removed.
Processing objective: The objective is to follow one benefit calculation from start to finish. The letter states the gross monthly benefit, takes the Medicare premium off it and states the payment that results. Each amount grounds to the sentence that holds it, along with the month the amounts took effect.
Demonstration fields: Benefit amounts effective from, Monthly benefit before deductions, Medicare premium deduction, Net monthly payment
The document
2 of 2 pages
Page 1shown below
Page 2
Where each value came from
current_benefits.period_start
{
"value": "December 2025",
"page": 1,
"range": {
"end": 511,
"start": 435
},
"box": {
"xmin": 0.11371,
"ymin": 0.45249,
"xmax": 0.82556,
"ymax": 0.47071
}
}current_benefits.gross_monthly_benefit
{
"value": 1137.3,
"page": 1,
"range": {
"end": 536,
"start": 512
},
"box": {
"xmin": 0.11418,
"ymin": 0.46935,
"xmax": 0.33626,
"ymax": 0.48576
}
}current_benefits.medical_insurance_premium_deduction
{
"value": 81.9,
"page": 1,
"range": {
"end": 598,
"start": 539
},
"box": {
"xmin": 0.11582,
"ymin": 0.49715,
"xmax": 0.69755,
"ymax": 0.51947
}
}current_benefits.net_monthly_payment
{
"value": 1055,
"page": 1,
"range": {
"end": 658,
"start": 601
},
"box": {
"xmin": 0.11512,
"ymin": 0.52881,
"xmax": 0.65211,
"ymax": 0.54565
}
}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.
