Document type

Extracting data from HELOC statements

A home equity line of credit revolves against the value of a house, and every lender lays its statement out differently. An underwriter reads one whenever an applicant already owns property, because the line is a liability against their income. Large lenders see thousands a month, arriving as portal downloads and scans.

An underwriter needs the borrowing limit, the outstanding balance, the available credit and the statement date. A debt-to-income calculation uses the balance and the payment, and the limit tells an automated underwriting system how much more could be drawn.

Start for Free Talk to Us

Page 1 of a sample HELOC Statement, with extracted values outlined

12,682Markdown characters

381parse blocks

168fields extracted

The problem

What makes HELOC statements hard

The balance runs backwards

The running balance counts borrowings, so money paid in makes it smaller and money drawn out makes it larger.

Four ledgers on one statement

A main account table and three sub-account tables each run their own rows, with interest moving between them.

Banking and lending in one

Pay goes in and bills go out of the same balance that carries the mortgage debt, so both kinds of row sit together.

How it works

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.

Worked example

One HELOC statement, end to end

About this document: A Manulife Bank sample statement for an all-in-one account: a revolving line secured on the home, combined with fixed and variable rate sub-accounts. Five portrait pages, marked for illustrative purposes only on every one, so no redaction was needed. Page 1 is featured, carrying the summary the rest of the statement details.

Processing objective: The objective is to read the three figures a lender compares: what may be borrowed, what is owed and what remains. They reconcile, since the limit less the outstanding balance gives the available credit to the cent.

Demonstration fields: Institution, Maximum borrowing limit, Available credit, Statement date, Outstanding balance

The complete example is on GitHubThe sample document, the full extraction schema, the complete parse output and every extracted field.

Open the folder

The document

4 of 5 pages

Page 1shown below
Page 2
Page 3
Page 4

Where each value came from

Page 1 of the sample HELOC Statement, with every featured field outlined

Institutionbank_name

Manulife Bank

The region of the sample HELOC Statement containing Institution, with the extracted value outlined

extraction_metadata["bank_name"]

{
  "value": "Manulife Bank",
  "page": 1,
  "range": {
    "end": 17,
    "start": 0
  },
  "box": {
    "xmin": 0.05272,
    "ymin": 0.04219,
    "xmax": 0.27852,
    "ymax": 0.06742
  }
}

Maximum borrowing limitsnapshot.borrowing_limit

240,000

The region of the sample HELOC Statement containing Maximum borrowing limit, with the extracted value outlined

extraction_metadata["snapshot.borrowing_limit"]

{
  "value": 240000,
  "page": 1,
  "range": {
    "end": 927,
    "start": 916
  },
  "box": {
    "xmin": 0.44729,
    "ymin": 0.49596,
    "xmax": 0.94349,
    "ymax": 0.52891
  }
}

Available creditsnapshot.available_limit

85,853.51

The region of the sample HELOC Statement containing Available credit, with the extracted value outlined

extraction_metadata["snapshot.available_limit"]

{
  "value": 85853.51,
  "page": 1,
  "range": {
    "end": 998,
    "start": 988
  },
  "box": {
    "xmin": 0.44735,
    "ymin": 0.52885,
    "xmax": 0.94362,
    "ymax": 0.566
  }
}

Statement datestatement_date

2023-05-01

The region of the sample HELOC Statement containing Statement date, with the extracted value outlined

Printed on the page as Prepared on May 1, 2023, normalized on extraction.

extraction_metadata["statement_date"]

{
  "value": "2023-05-01",
  "page": 1,
  "range": {
    "end": 129,
    "start": 106
  },
  "box": {
    "xmin": 0.05484,
    "ymin": 0.15159,
    "xmax": 0.21151,
    "ymax": 0.16708
  }
}

Outstanding balancesnapshot.closing_total_borrowings

154,146.49

The region of the sample HELOC Statement containing Outstanding balance, with the extracted value outlined

extraction_metadata["snapshot.closing_total_borrowings"]

{
  "value": 154146.49,
  "page": 1,
  "range": {
    "end": 691,
    "start": 680
  },
  "box": {
    "xmin": 0.44717,
    "ymin": 0.42314,
    "xmax": 0.94319,
    "ymax": 0.44284
  }
}

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

Run this sample yourself This document on GitHub

Why LandingAI?

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.

Compare plans

Capabilities

HELOC Statement processing highlights

Every lender's layout

Revolving credit limits

Sub-account ledgers

Fixed and variable rates

Interest between sub-accounts

Full transaction ledgers

Available credit figures

Statement period dates

Canadian and US statements

Dates normalized to ISO

Bounding box for every value

Zero data retention available

FAQs

Questions, answered

Every sub-account. Model each as its own object and its rows, rate and balance come back attributed to it, so a fixed rate figure never lands in the variable rate one.

Yes. Each row comes back with its amount and the column it sat in, so the direction is read from what the statement prints, and the running balance beside it can be checked against that.

No. One schema covers every lender. A field a particular statement does not carry comes back null, and the rest come back under the names you gave them.

Every value comes back with its page, a bounding box and the character range in the parsed Markdown. A date returned in ISO form still points at the date the statement printed, so a reviewer sees both.

A HELOC statement carries a borrower, an account number, a limit secured on their home and a full transaction ledger. A schema can ask for the summary figures alone. Accounts are available with zero data retention, alongside SOC 2 Type II, GDPR and HIPAA compliance.

Start extracting HELOC statement data

Reliable, structured outputs with full traceability in minutes