Skip to main content
Document Verification enables businesses to confirm the authenticity of physical identity documents submitted by users, such as national IDs, passports, or driver’s licenses. This is achieved by analyzing uploaded images and extracting key information. The system uses a combination of image analysis, document template matching, optical character recognition (OCR), and fraud signals to determine whether the document is valid, tampered with, or incorrectly submitted. This service is built to support onboarding, KYC, and risk checks across sectors such as fintech, lending, crypto, insurance, and mobility.

​
How it Works

The document verification process typically follows these steps:
1

Image Upload

The user uploads a photo or scanned image of their document, usually the front, and optionally the back side. Images can be uploaded via API, SDK, or manually through the dashboard.
2

Preprocessing and Image Quality Check

The system performs checks for:
  • Blurriness
  • Glare or poor lighting
  • Incomplete edges or cropped content
  • Incorrect document orientation or rotation
  • Document legibility
If the image fails basic quality rules, the system returns a failed response with details for retry.
3

Template Matching

The system determines the document type and matches it against our library of over 11,000+ document templates across 200+ countries. This helps identify if the document structure aligns with expected formats. You can view documents here.
4

OCR and Data Extraction

Key data fields are extracted from the document using OCR. This includes:
  • Full name
  • Date of birth
  • Document number
  • Issuance and expiry dates
  • Nationality or issuing country
  • MRZ (for passports), amongst others.
5

Authenticity and Fraud Check

We evaluate the document for signs of forgery or manipulation. This includes:
  • Font inconsistencies
  • Layer tampering
  • Resized or resubmitted document images
  • Edited MRZ or altered expiry dates
If enabled, you can also run additional checks like:
  • Face match with a selfie
  • Cross-checks with government lookup sources

​
How to Run a Document Analysis

You can implement Document Verification in three ways:

API Lookup

Send a POST request with the document image and metadata. Receive structured output with extracted fields, fraud risk signals, and image quality score.

SDK/Libraries

Use our SDK to build the full document capture and verification flow into your product. This Includes auto-capture with camera guidance, real-time feedback on image quality, and optional selfie capture for face match

Manual Lookup via Dashboard

Need something quick? The dashboard lets you run one-off identity checks without writing a single line of code. Just upload and verify documents manually through the Dojah dashboard.

​
Supported documents

We support government-issued identity documents from over 200+ countries, including:
  • National ID cards
  • Passports
  • Driver’s licenses
  • Voter cards
  • Residence permits
  • Work authorization cards, etc
Each document type is validated against its unique structure and layout, including MRZ where applicable. See full list of supported documents.