Extracting data from SSA-1099 statements
The Social Security Administration uses SSA-1099 to report annual Social Security benefits on a federal layout. Tax preparers and lenders may request the form when reporting or verifying that income. Recipients can provide a scan, photo or download from their online account.
Lenders and preparers need the benefits paid, the benefits repaid, the net benefits and any tax withheld. The loan origination system grosses up the nontaxable portion as qualifying income, and tax software carries the net benefits onto the return.
What makes SSA-1099 statements hard
Boxes that refer to each other
Box 5 is box 3 minus box 4, so the net benefit and the gross benefit can print the same figure in adjacent boxes.
Labels inside the amounts
Each box prints its label and amount together, such as benefits paid in the year followed by the dollar figure, on one line.
Detail behind the totals
The description of amounts lists what makes up box 3 and box 4, such as Medicare premiums deducted, beneath the headline boxes.
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 SSA-1099, end to end
About this document: An SSA-1099 from the IRS certification test booklet for volunteer tax preparers, which prints filled specimen tax documents for invented taxpayers. One landscape page cut from the booklet, with its numbered boxes and the description of amounts. Page 1 is featured. The beneficiary is fictitious and the identifier is already masked.
Processing objective: The objective is to read the year's benefit summary: which year it covers, what was paid, the net after any repayment, and the tax withheld. Each value grounds to the numbered box that prints it.
Demonstration fields: Benefit year, Benefits paid (box 3), Net benefits (box 5), Federal tax withheld (box 6)
The document
1 page
Page 1shown below
Where each value came from
tax_year
{
"value": "2026",
"page": 1,
"range": {
"end": 60,
"start": 56
},
"box": {
"xmin": 0.08444,
"ymin": 0.14198,
"xmax": 0.18056,
"ymax": 0.18481
}
}box_3_benefits_paid
{
"value": 25182,
"page": 1,
"range": {
"end": 375,
"start": 336
},
"box": {
"xmin": 0.08718,
"ymin": 0.23606,
"xmax": 0.25933,
"ymax": 0.29123
}
}box_5_net_benefits
{
"value": 25182,
"page": 1,
"range": {
"end": 501,
"start": 442
},
"box": {
"xmin": 0.48748,
"ymin": 0.23683,
"xmax": 0.76549,
"ymax": 0.29094
}
}box_6_federal_tax_withheld
{
"value": 2518,
"page": 1,
"range": {
"end": 868,
"start": 811
},
"box": {
"xmin": 0.42626,
"ymin": 0.50877,
"xmax": 0.76554,
"ymax": 0.58666
}
}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.
