KYC & onboarding
Build a reusable document-data extraction layer into your onboarding stack.
IdScanly reads supported identity-document images and returns structured fields your backend can map, review and pass to the next step. Process the front, back and retakes of the same document in one session, with confidence scores and capture feedback to support your user journey.
Synthetic data. A possible integration, not a ready-made connector or a live product screen.
DEMO-0001Document images contain information your application needs in a different form: fields for an application record, values for comparison and inputs for downstream checks.
A document OCR API converts those images into structured output. IdScanly provides that extraction layer for teams building customer registration and KYC workflows across web, mobile and staff-assisted channels.
Your application controls the interface, required documents, review policy and connections to other services. The extraction component supplies document-derived data without requiring you to build the reading pipeline yourself.
Identity information may be split across the front and back of a card. A passport image may contain one clear section and another obscured by glare. Asking for a retake should help the same document flow continue.
With IdScanly, keep the sides and retakes of the same document in one scan session. Upload the first image when it is ready, then add further views while processing continues. The service combines available evidence into a merged result.
Use separate sessions for different documents, such as a passport and a driver license. Your own onboarding record connects those sessions to the correct person and application.
Your backend can start the session with the first image instead of waiting for every required view. Your interface can show the captured image so the customer can check framing, lighting and readability while continuing the document step.
When scan feedback identifies an image issue, show a specific next action. A customer may need to retake a blurred image, include a missing edge or provide another view of the document.
Partial fields can become available during asynchronous processing. Your application can use these to prepare the interface while keeping processing status and final confirmation distinct.
The final result contains combined fields and confidence information. Apply your own rules for confirmation, another capture or human review, particularly for required fields and meaningful differences from declared data.
Depending on the supported document, IdScanly can return personal information, document identifiers, issuing details, validity dates, printed addresses and document-specific information.
Examples include names, date and place of birth, nationality, document number, issuing country and expiry date. Driver licenses may also contain supported categories or restrictions; other document types expose different fields.
MRZ reading with checksum validation and supported barcode extraction can provide additional document data. AAMVA PDF417 data is supported on relevant US and Canadian driver licenses. Available output depends on the document format, version and image quality.
Design your mapping around the fields your workflow needs, while preserving the distinction between a field that is unavailable and one that has been reviewed and confirmed.
Keep API credentials server-side. Your backend should also handle request failures, interrupted customer journeys and result retrieval according to the current API documentation.
Explore the API documentationOCR extracts information printed or encoded in a document image. Full identity verification can also require evidence of document authenticity, a connection between the presenter and the document, liveness and other checks relevant to the business.
IdScanly supplies the extraction component. Your workflow chooses and coordinates the additional services it needs. A confidence score concerns extraction quality, while MRZ checksum validation concerns encoded-data consistency; neither independently verifies the customer’s identity.
This division gives your team control over the wider onboarding architecture and the data each component receives.
Build a representative test set covering your expected countries, document types and image conditions. Compare field-level output, including names with diacritics, document numbers, dates, missing fields and results combined from several images.
Test the user journey as well as the successful scan. Check what your interface does after an unclear photo, an omitted side, an interrupted upload or a result requiring review.
For larger processing volumes, discuss dedicated-server capacity and support requirements with IdScanly. Review current processing and privacy terms, including temporary data involved in processing and result retrieval.
Document OCR extracts data from document images. KYC is a broader customer due-diligence workflow that can require identity verification, screening, risk assessment and ongoing controls. OCR can provide an input to that workflow.
Yes. Keep images of the same document in one scan session so their readable information can contribute to the merged result.
Partial results can be available during processing. Your application should distinguish those values from the completed result and manage when the user can confirm the information.
No. It supports assessment of the extracted reading. Authenticity and holder verification require separate capabilities.
Keep it unresolved and use your capture, correction or review process. Your application should not invent a value or interpret missing output as confirmed information.
No. The extraction layer can also support guest registration, rental administration, visitor management and other document-based onboarding workflows.
Try IdScanly with the documents your customers use, then review the API workflow with your engineering team. Share the fields, document coverage and processing volumes your product needs.
Explore business workflows: Fintech applications · Car rental · Hotel check-in