Extracting data from odometer disclosure statements
Federal law requires a signed mileage statement when most vehicles change hands, and states and lessors each print their own form for it. An auto lender's funding team checks one with each used vehicle deal whose title has no odometer section. They arrive as scans and phone photos uploaded through a dealer portal.
Funding review needs the odometer reading, the date it was read, the VIN and the certification the seller ticked. A loan origination system checks the mileage against the vehicle value behind the loan amount, and flags a reading certified as other than the actual mileage.
What makes odometer disclosure statements hard
Mileage split across digit boxes
Most forms print the reading as a row of single-digit boxes, broken by a printed comma and a tenths position.
The certification is a checkbox
The seller certifies the reading by ticking one of three printed statements: actual mileage, beyond the odometer's mechanical limits, or not the actual mileage.
One disclosure, many forms
The federal disclosure appears in a title's odometer section, on state separate forms and on lessor statements, each with its own layout.
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 odometer disclosure statement, end to end
About this document: Florida's Separate Odometer Disclosure Statement and Acknowledgment, form HSMV 82993 from the Department of Highway Safety and Motor Vehicles. Instafill.ai filled it in for a private sale and published it as a worked example. One page, a fillable PDF, carrying a vehicle description table, the odometer statement in spaced digit boxes, three certification checkboxes, and seller and buyer blocks. Both names and street addresses were replaced with stand-ins.
Processing objective: The objective is to capture one complete disclosure: which vehicle, what its odometer read, on what date, and what the seller certified about that reading. The reading is assembled from its digit boxes into a whole number. The certification comes back as the wording of the ticked statement.
Demonstration fields: Vehicle year, Vehicle make, Odometer reading (miles), Date read, Mileage certification
The document
1 page
Page 1shown below
Where each value came from
vehicle.year
{
"value": 2018,
"page": 1,
"range": {
"end": 371,
"start": 358
},
"box": {
"xmin": 0.47052,
"ymin": 0.15588,
"xmax": 0.53683,
"ymax": 0.19318
}
}vehicle.make
{
"value": "Honda",
"page": 1,
"range": {
"end": 393,
"start": 380
},
"box": {
"xmin": 0.53656,
"ymin": 0.15588,
"xmax": 0.60294,
"ymax": 0.19318
}
}odometer.reading_miles
{
"value": 75123,
"page": 1,
"range": {
"end": 851,
"start": 758
},
"box": {
"xmin": 0.08752,
"ymin": 0.27012,
"xmax": 0.87882,
"ymax": 0.28973
}
}odometer.date_read
Printed on the page as DATE READ 03 / 15 / 2026, AND WE HEREBY CERTIFY THAT TO THE BEST OF OUR KNOWLEDGE THE, normalized on extraction.
{
"value": "2026-03-15",
"page": 1,
"range": {
"end": 938,
"start": 852
},
"box": {
"xmin": 0.08823,
"ymin": 0.28969,
"xmax": 0.82568,
"ymax": 0.30777
}
}mileage_certification
{
"value": "REFLECTS ACTUAL MILEAGE",
"page": 1,
"range": {
"end": 1040,
"start": 1008
},
"box": {
"xmin": 0.24671,
"ymin": 0.32342,
"xmax": 0.48239,
"ymax": 0.34121
}
}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. VPC and on-prem deployment require an Enterprise plan.
