1

Zera AI — The Core Engine

The Zera AI engine is what separates Zera Books from converters built on generic PDF parsing. It was trained on 3.2 million financial documents: 2.8 million bank statements from hundreds of institutions, 420,000 invoices, and a combined 847 million individual transactions.

That training volume matters because bank statement formats vary enormously. Chase formats its statements differently from RBC, TD, PNC, and Scotiabank. Within a single bank, the format can shift when the institution updates its PDF template. A tool built on rules and templates breaks at that point. Zera AI identifies column structure, header patterns, and transaction rows dynamically — it doesn't rely on a fixed map.

Zera AI — Training & Accuracy Stats

Documents trained on
3.2M+
Bank statements in training set
2.8M
Transactions analyzed
847M
Field-level accuracy
99.6%
2

Document Types and Input Formats

Most document processing tools are built around one use case and bolted onto others. Zera Books was architected from the start to handle four document types — because that's what accounting workflows actually require. You don't process bank statements in isolation; you're reconciling them against invoices and checking them against financial reports.

Document TypeFields ExtractedAccepted Input
Bank StatementsDate, description, debit, credit, running balance, account number, opening/closing balanceDigital PDF, scanned PDF, JPG, PNG, multi-page, password-protected
Financial StatementsLine items, period headers, account codes, subtotals, comparative column dataDigital PDF, scanned PDF
InvoicesVendor name, invoice date, invoice number, line items, quantities, unit prices, tax amounts, PO number, payment termsDigital PDF, scanned PDF, JPG, PNG
ChecksMICR routing number, account number, check number, payee, amount in numerals and text, date, memo fieldScanned PDF, JPG, PNG

Batch input: Upload 50+ documents simultaneously. The platform processes them in parallel and returns output files organized by document type and client. See client management reference for dashboard organization details.

See the platform in action

One-week trial. No template setup, no per-page fees. Process your first statement in under 60 seconds.

Try for one week
3

Integrations and Output Formats

There are two tiers of integration: direct API connections where data flows automatically into your accounting software, and pre-formatted exports where the file is structured to match the import requirements of the target system.

QuickBooks Online — Direct API

Categorized transactions pushed directly into QBO. No export/import cycle, no manual field mapping. Includes auto-categorization mapped to your existing chart of accounts.

Xero — Direct API

Same direct pipeline for Xero. Transactions categorized against your Xero chart of accounts with confidence scoring. Review and approve before import.

Pre-Formatted Exports

Sage, Wave, Zoho, NetSuite, FreshBooks, MYOB, Oracle, IIF (QuickBooks Desktop). Each export matches the exact column structure required for a clean import. No manual mapping needed.

Universal Export Formats

Excel (XLSX), CSV, QBO, IIF. All include auto-categorized transaction data. PDF to Excel specs and PDF to CSV specs on separate reference pages.

4

Workflow Features

The platform features beyond extraction are what make Zera Books suitable for practices with 10+ clients. Single-document converters don't scale — you need organization, history, and batch capability built in.

1

Multi-account auto-detection

A single PDF containing checking, savings, and credit card data is automatically identified and split into separate output files. No manual page-range selection.

Detects account boundaries using header pattern recognition — works on statements that mix multiple accounts on the same page. Reference: multi-account detection page.
2

Duplicate detection

Overlapping statement periods are common when clients provide statements that cover the same date range from multiple sources. The deduplication engine compares transaction date, amount, and description to flag and remove duplicates before export.

3

Data standardization

Date formats vary by bank and region (MM/DD/YYYY vs DD/MM/YYYY vs MMM DD, YYYY). Description text contains bank-specific reference codes and truncations. The platform normalizes both to a consistent format across all output files.

4

Audit trail

Every conversion is logged with timestamp, source file name, user account, and output files generated. Unlimited conversion history — useful for firms with regulatory record-keeping requirements.

5

Security Architecture

Financial documents contain sensitive client data. The Zera Books security model is built around three principles: encryption at every layer, minimal data retention, and explicit permission before anything is stored beyond the processing window.

AES-256 Encryption

All data encrypted in transit (TLS 1.3) and at rest (AES-256). No plaintext storage at any layer of the stack.

30-Day Auto-Delete

Uploaded files are automatically deleted after 30 days. No data retained beyond that window without explicit account holder permission.

No Sharing Without Permission

Uploaded documents and extracted data are not used to train models or shared with third parties without explicit opt-in.

Conversion Audit Log

Every document processed is logged with timestamp and account identifier. Supports internal compliance reviews and client record-keeping requirements.

6

Frequently Asked Questions

What document types does Zera Books process?

Four types: bank statements, financial statements (P&L, balance sheets, cash flow), invoices with line-item extraction, and checks with MICR data. Most competitors process only bank statements.

Does Zera Books require template setup for new banks?

No. Zera AI identifies document structure dynamically. It adapts to format changes and previously unseen layouts without any configuration. See Zera AI reference for model details.

What accounting software does Zera Books connect to?

Direct API integration with QuickBooks Online and Xero. Pre-formatted exports for Sage, Wave, Zoho, NetSuite, FreshBooks, MYOB, and Oracle. All exports include AI-categorized transactions with no manual field mapping.

How is client data protected?

AES-256 encryption in transit and at rest. Files auto-deleted after 30 days. No data sharing without explicit permission. Full details at zerabooks.com/platform.