use case

Identity checks for customer onboarding

Turn a photo of an ID and a selfie into checks your team can review, with every result and its evidence in one place.

the problem

What teams deal with

Typing details by hand slows customers down and introduces errors.
Photos arrive blurred, cropped or edited, and need flagging rather than guessing.
Aadhaar numbers must be masked and never kept unmasked.

a typical flow

Each step is a separate API with the same request and response design. Your policy decides which steps a product needs.

apis

PAN card OCR API
Read the PAN, name and other printed details from an uploaded PAN card image, with a confidence score and warnings for each field.
PAN verification API
Check that a PAN exists and is active, and see the permitted details on record, with the source and time of every field.
Aadhaar OCR and masking API
Read the permitted details from an Aadhaar card image and mask the Aadhaar number, which is never stored unmasked.
Face match and liveness API
Compare a selfie with the photo on an identity document, and check a capture for signs of a presentation attack, as two separate results.
Document tampering detection API
Flag signs that an ID or supporting document was edited, with the affected region, severity and confidence for each indicator.
PAN name and date-of-birth match API
Compare the name and date of birth an applicant entered with the permitted PAN details, field by field, with a reason for every outcome.

good to know

Your KYC obligations remain yours. Koshio supplies checks and the evidence behind them; your policies decide what happens next.
Biometric results are observations with stated limits, never a verdict that a person is who they claim to be.

questions

Is the Aadhaar number stored?

Never unmasked. The Aadhaar OCR and masking API returns the number masked to its last four digits and does not keep the original.

Tell us what your flow looks like today, and we'll show you how the checks fit into it.