Extracting data from consolidated 1099s
A brokerage sends one consolidated 1099 to every client with an active account, bundling five IRS forms into a single statement. Fidelity, Schwab, Vanguard and Edward Jones each lay theirs out differently. A tax preparer works through hundreds in a six-week season, arriving as portal downloads and posted paper.
A preparer needs every 1099-B transaction: the security, the acquisition and sale dates, the proceeds, the cost basis and the resulting gain. Tax software totals those into Schedule D, and reconciles them against the statement's own summary boxes.
What makes consolidated 1099s hard
Many forms in one document
A single statement carries five IRS forms, each with its own box numbering. Extraction has to know which section a figure belongs to.
Transaction detail runs for pages
An active account produces hundreds of 1099-B rows across many pages, grouped into short-term and long-term, covered and noncovered sections.
Every broker formats differently
Box ordering, section naming and page layout vary by issuer. Any vendor solution relying on position-based extraction is rebuilt per broker, every tax year.
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 consolidated 1099, end to end
About this document: A Consolidated 1099 Statement from Edward Jones, 13 landscape pages. It bundles 1099-B, 1099-INT, 1099-DIV, 1099-OID and 1099-MISC, with pages of supplemental detail the IRS does not require. Page 4 carries the long-term 1099-B transactions. The recipient name, account number and advisor details were replaced with fictitious equivalents before publication, and the page carrying the mailing address was removed.
Processing objective: The objective is to lift one complete long-term 1099-B transaction off page 4, with each figure traceable to its own cell. The row chosen reconciles: proceeds of 3,071.56 less a cost basis of 2,982.67 gives the 88.89 gain.
Demonstration fields: Security sold, Date acquired, Gross proceeds, Cost basis, Gain or loss
The document
4 of 13 pages
Page 1
Page 4shown below
Page 5
Page 8
Where each value came from
form_1099_b.details[1].description_of_property
{
"value": "AMERICAN CAP INC BUILDER A / 140193103 / CAIBX",
"page": 4,
"range": {
"end": 10942,
"start": 10896
},
"box": {
"xmin": 0.04563,
"ymin": 0.57507,
"xmax": 0.95501,
"ymax": 0.5976
}
}form_1099_b.details[1].date_acquired
Printed on the page as 09/02/2023, normalized on extraction.
{
"value": "2023-09-02",
"page": 4,
"range": {
"end": 11001,
"start": 10991
},
"box": {
"xmin": 0.24673,
"ymin": 0.59753,
"xmax": 0.31505,
"ymax": 0.61772
}
}form_1099_b.details[1].proceeds
{
"value": 3071.56,
"page": 4,
"range": {
"end": 11037,
"start": 11029
},
"box": {
"xmin": 0.38291,
"ymin": 0.59754,
"xmax": 0.5006,
"ymax": 0.61774
}
}form_1099_b.details[1].cost_basis
{
"value": 2982.67,
"page": 4,
"range": {
"end": 11054,
"start": 11046
},
"box": {
"xmin": 0.50057,
"ymin": 0.59755,
"xmax": 0.59613,
"ymax": 0.61775
}
}form_1099_b.details[1].gain_or_loss
{
"value": 88.89,
"page": 4,
"range": {
"end": 11068,
"start": 11063
},
"box": {
"xmin": 0.59611,
"ymin": 0.59756,
"xmax": 0.68701,
"ymax": 0.61776
}
}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.
