Fields Extracted from Checks
A check contains two distinct data zones: the printed fields in the body (payee, amount, date, memo) and the MICR line at the bottom (machine-readable bank routing and account identifiers). Zera Books extracts both zones in a single pass.
| Field | Location on Check | Extraction Notes |
|---|---|---|
| Routing Number | MICR line — first 9 digits | E-13B and CMC-7 MICR fonts handled. 97%+ accuracy on clean scans. |
| Account Number | MICR line — second field | Varies in length (8–17 digits). Extracted as string to preserve leading zeros. |
| Check Number | MICR line and upper right corner | Cross-validated between MICR and printed check number. |
| Payee Name | "Pay to the Order of" line | Full name extracted; handles handwritten payee names at 92%+ accuracy. |
| Amount (Numeral) | Amount box (upper right area) | Numeric value extracted; currency symbols stripped. Validated against written amount. |
| Amount (Written) | "______ Dollars" line | Extracted for validation against numeric amount. Discrepancies flagged in output. |
| Check Date | Date field (upper right) | Normalized to YYYY-MM-DD. Handles handwritten dates including abbreviated month names. |
| Memo / Reference | Memo line (lower left) | Full text extracted where legible. Common for PO references and invoice numbers. |
| Bank Name | Bank logo / header area | Extracted from printed bank name; used to validate routing number. |
MICR Line Processing
MICR (Magnetic Ink Character Recognition) uses specialized fonts — E-13B in North America and CMC-7 in Europe — designed to be read by magnetic scanners. While checks are typically scanned optically rather than magnetically in digital workflows, the MICR font characters are sufficiently distinct that a well-trained OCR model can extract them reliably from scanned images.
Zera OCR includes MICR-specific training data for both E-13B and CMC-7 character sets. The processing pipeline applies additional validation: routing numbers are checked against the American Bankers Association routing number format, and account numbers are cross-referenced against the check number extracted from the upper right corner of the check.
MICR Extraction — Technical Specs
Process check batches in minutes, not hours
Upload 50+ scanned check images. Get MICR data, payee, amount, and date extracted in structured format — reconciliation-ready. Included in $79/month unlimited.
Try for one weekCheck Data in Reconciliation Workflows
The value of check data extraction is in reconciliation — matching checks issued against checks that have cleared the bank. Here's how extracted check data integrates with the reconciliation workflow:
Check Register vs Bank Statement
Extracted check data (number, amount, payee, date) can be cross-referenced against extracted bank statement transactions using check number as the matching key.
Outstanding Check Identification
Checks in the extracted register that don't appear as cleared in the bank statement data are flagged as outstanding — a key step in the monthly bank reconciliation.
Amount Discrepancy Detection
When the numeric and written amounts on a check don't match, the discrepancy is flagged in the output before the data is used in reconciliation — catching potential fraud or errors.
AP Matching
Memo field content (PO numbers, invoice references) extracted from checks enables matching against open payables — useful for confirming payment of specific invoices.
Full reconciliation workflow: Zera Books processes both bank statements and checks. When both are uploaded for the same period, the platform can cross-reference check numbers between the bank statement transactions and the extracted check register. See bank reconciliation reference for the full workflow.
Frequently Asked Questions
What data does Zera Books extract from checks?
MICR line (routing number, account number, check number), payee name, amount in numerals and words, check date, memo field, and bank name. All fields extracted as separate structured columns.
Can Zera Books read the MICR line from scanned checks?
Yes. Zera OCR is trained on MICR font (E-13B and CMC-7) and achieves 97%+ extraction accuracy on clean scans at 200 DPI or above. Routing numbers are validated against ABA format.
How does check processing integrate with bank reconciliation?
Extracted check data is structured with the same field format as bank statement transactions, allowing check number-based cross-referencing during reconciliation. See bank reconciliation reference for details.
What input formats are accepted for check processing?
Scanned PDFs, JPG, and PNG images. Input checks at 150 DPI minimum; 200+ DPI recommended for best MICR accuracy. Batch upload of 50+ check images supported. Full details at zerabooks.com/products/checks.