Document OCR reads identity and business documents from photos or PDFs, extracts the fields, validates checksums, and flags tampering. It's tuned for the documents Bangladeshi businesses actually see — not a generic Western OCR repackaged for the region.
Supported document types
- National ID (NID) — old paper card, smart card, both sides.
- Passport — MRZ parsing, validity check, photo extraction.
- Driving licence — BRTA-issued cards.
- Trade licence — city corporation and union parishad formats.
- TIN / BIN certificate — NBR formats with checksum validation.
Tamper detection
We score every document on: font consistency, edge integrity, hologram presence (where applicable), photo resampling artifacts, and metadata anomalies. Tampered submissions are flagged for manual review with a heat-map highlighting the suspect region.
Used in corporate banking onboarding and seller verification.