Document type

Extracting data from passports

Every country designs its own passport, and an ICAO standard fixes the machine-readable zone at its foot. An airline boarding a passenger and a bank opening an account both read one to confirm who someone is. A large onboarding team checks thousands a month, arriving as phone photographs through an app.

A verifier needs the passport number, the issuing country, the date of expiry and the two machine-readable lines. An identity platform recomputes the MRZ check digits against the printed fields, then screens the holder against sanctions lists.

Start for Free Talk to Us

Page 1 of a sample Passport, with extracted values outlined

606Markdown characters

1parse blocks

10fields extracted

The problem

What makes passports hard

The MRZ repeats everything

Two lines of fixed-format text at the foot of the page carry the identity data again, with check digits over each group.

Labels come in two scripts

A passport labels each field in its own language and again in a second, and carries fields particular to the country that issued it.

Curved pages and laminate glare

The identity page is photographed open, so it curves at the spine and the laminate throws glare across the print.

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 passport, end to end

About this document: The official specimen of the identity page of Japan's passport, published by the Ministry of Foreign Affairs. One landscape image, bilingual throughout, with a two-line machine-readable zone along the foot and a handwritten signature in kanji and hiragana. Every value is specimen data, with SPECIMEN printed across the page, so no redaction was needed.

Processing objective: The objective is to read the document's identity off a single image: what it is, who issued it and when it was issued. The issue date is returned as an ISO date, normalized from the day-month-year the page prints.

Demonstration fields: Document type, Passport number, Issuing country, Sex, Date of issue

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

1 page

Page 1shown below

Where each value came from

Document typedocument_type

Passport

extraction_metadata["document_type"]

{
  "value": "Passport",
  "page": 1,
  "range": {
    "end": 46,
    "start": 38
  },
  "box": {
    "xmin": 0.16032,
    "ymin": 0.21425,
    "xmax": 0.29217,
    "ymax": 0.25035
  }
}

Passport numberdocument_info.document_number

ZZ0000826

extraction_metadata["document_info.document_number"]

{
  "value": "ZZ0000826",
  "page": 1,
  "range": {
    "end": 114,
    "start": 105
  },
  "box": {
    "xmin": 0.66971,
    "ymin": 0.2445,
    "xmax": 0.84877,
    "ymax": 0.28703
  }
}

Issuing countrydocument_info.issuing_country

JAPAN

extraction_metadata["document_info.issuing_country"]

{
  "value": "JAPAN",
  "page": 1,
  "range": {
    "end": 9,
    "start": 4
  },
  "box": {
    "xmin": 0.70137,
    "ymin": 0.15423,
    "xmax": 0.83184,
    "ymax": 0.19791
  }
}

Sexpersonal_info.sex

extraction_metadata["personal_info.sex"]

{
  "value": "F",
  "page": 1,
  "range": {
    "end": 205,
    "start": 204
  },
  "box": {
    "xmin": 0.33046,
    "ymin": 0.52174,
    "xmax": 0.35816,
    "ymax": 0.5586
  }
}

Date of issuedocument_info.date_of_issue

2025-03-24

Printed on the page as 24 MAR 2025, normalized on extraction.

extraction_metadata["document_info.date_of_issue"]

{
  "value": "2025-03-24",
  "page": 1,
  "range": {
    "end": 320,
    "start": 309
  },
  "box": {
    "xmin": 0.3312,
    "ymin": 0.58968,
    "xmax": 0.56773,
    "ymax": 0.62824
  }
}

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

Passport processing highlights

Every country's passport design

Machine-readable zone lines

Bilingual field labels

Non-Latin scripts

Handwritten signatures

Passport and document numbers

Issue and expiry dates

Nationality and issuing country

Photographed identity pages

Country-specific fields

Dates normalized to ISO

Bounding box for every value

FAQs

Questions, answered

Yes. Both MRZ lines come back character for character, including the filler characters, so the check digits can be recomputed and compared against the printed fields above them.

Yes. Labels and values in a non-Latin script parse as text, alongside the Latin transliteration a passport prints beside them. A handwritten signature comes back as text too.

No. One schema covers passports from any issuing authority. A field a particular country does not print comes back null, and the fields it does print come back attributed to 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 day-month-year printed on the page, so a reviewer sees both.

A passport carries a name, a date of birth, a document number and a portrait. A schema can ask for the document fields alone. Accounts are available with zero data retention, alongside SOC 2 Type II, GDPR and HIPAA compliance.

Start extracting passport data

Reliable, structured outputs with full traceability in minutes