01 · HANDWRITINGlive demo
Handwritten cheques, not just printed text.
Cursive payees, rushed signatures, stylised dates. Our field-specific models are optimised for and evaluated on real cheques.
98.4%field-level accuracy
cursive readyExtract MICR, amount, payee, date, and endorsement data from bank checks — including handwritten fields that generic OCR misses. Chequedb is purpose-built for cheque handwriting, legal amount reconciliation, and deployment inside your own infrastructure.
Use this page to evaluate fit for mobile deposit, RDC, teller capture, and back-office automation. Use the API page for integration detail and request shapes.
Three capabilities almost nobody else ships together — handwriting trained on real cheques, a parser that understands how people actually say money, and deployment that keeps every account number inside your network.
Cursive payees, rushed signatures, stylised dates. Our field-specific models are optimised for and evaluated on real cheques.
“Lakh”, “naira only”, “k”, “no cents”, “no paisa”. Our parser handles regional financial language and reconciles it against the numeric box every time.
Cheque images carry account numbers, signatures, and payment history. Deploy ChequeDB inside your infrastructure so the data never leaves your perimeter.
Generic cloud OCR can read text. Almost nobody combines cheque-specific handwriting recognition, colloquial amount reconciliation, and on-prem deployment for non-bank businesses. That’s the gap ChequeDB fills.
Book demoBank check OCR (Optical Character Recognition) extracts structured data from check images: routing number, account number, amount, payee name, issue date, endorsement text, and signature region. It combines two reading methods — MICR for the encoded control line at the bottom of the check, and OCR/ICR for the visible printed and handwritten fields above it.
MICR reads routing, account, and check number from the magnetic-ink line. OCR reads the numeric amount box. ICR handles handwritten fields like the legal amount line, payee, and date. A complete bank check OCR system uses all three, returns per-field confidence scores, and routes low-confidence fields to a review queue rather than silently auto-accepting them.
The extraction output feeds validation: amount cross-check, date rules (stale-dated and post-dated flags), routing validation, and duplicate detection. Each validation step produces a clear status so the downstream deposit or review workflow can act without rebuilding that logic. For a deeper technical walkthrough, see Bank Check OCR: What It Reads, How It Works, and How to Integrate It.
Capture the fields operations teams actually need for review, posting, and exception handling.
| Field Group | Data Returned | Why It Matters |
|---|---|---|
| MICR and bank data | Routing number, account number, check number, bank identifier | Supports posting, routing validation, duplicate checks, and downstream workflow matching. |
| Amount and payee | Courtesy amount, legal amount, payee name, memo region | Helps teams compare numeric and written values, match payee rules, and catch mismatch-driven exceptions. |
| Date and image quality | Issue date, stale or future-date flags, blur and framing checks | Prevents bad submissions from entering deposit operations and improves straight through processing rates. |
| Signature and endorsement | Signature region, endorsement presence, review-ready image coordinates | Gives fraud, operations, and reviewer tools the context needed for faster manual checks and approval routing. |
One check OCR layer for image ingestion, validation, decision support, and deposit handoff.
Accept front and back images from mobile apps, scanner-based RDC, teller desks, or kiosk flows.
Read MICR and visual fields, localize signatures and endorsements, and return confidence-aware JSON.
Score image quality, compare amounts, validate routing, flag duplicates, and route exceptions to reviewers.
Push accepted items into deposit and review workflows, then connect deeper implementation detail through the check processing API.
{
"check_id": "chk_01hy9v3x8c",
"status": "review_required",
"fields": {
"routing_number": { "value": "021000021", "confidence": 0.99 },
"account_number": { "value": "1234567890", "confidence": 0.98 },
"check_number": { "value": "1042", "confidence": 0.99 },
"amount": { "value": 1850.25, "confidence": 0.96 },
"payee": { "value": "Summit Supply LLC", "confidence": 0.87 },
"date": { "value": "2026-03-20", "valid": true }
},
"checks": {
"image_quality": "pass",
"duplicate_risk": "low",
"amount_match": "pass",
"endorsement_present": true
}
}Start with the delivery model that matches your application surface and control requirements.
Built for operations teams that need field-level confidence signals, exception traceability, and flexible deployment.
The same OCR layer across mobile, scanner, and back-office check capture.
Use the API for integration, then connect the same extracted fields to review, archive, and management workflows.
OCR software for extracting cheque fields before API handoff.
Route low-confidence OCR results into approval and exception queues.
Capture quality-controlled cheque images from scanners and mobile flows.
Store extracted fields, images, statuses, and audit history for search.
Book a demo to evaluate workflow fit. For request shapes, JSON response, and SDK options, move to the API docs.