Accounts Payable Parser API

The API supports PDF, DOCX, and plain text formats for invoice processing. PDF support includes both digitally-generated invoices (with embedded text layers) and scanned paper invoices processed through integrated OCR with automatic image enhancement. DOCX parsing handles complex document structures including nested tables, merged cells, and multi-column layouts commonly used in vendor invoices. The plain text endpoint (/v1/parse/text) accepts JSON or raw text payloads, making it ideal for invoices received via email body text or EDI systems. The maximum file size for uploads is 10MB, and the API automatically detects document orientation and applies rotation correction for scanned documents.

The text parsing endpoint (/v1/parse/text) accepts invoice content as either a JSON payload with a "text" field or as raw text with a text/plain content type. The AI engine applies the same extraction models used for document parsing, identifying vendor information, invoice details, line items, and payment terms from unstructured text. This endpoint is particularly useful for processing invoices received as email body content, EDI 810 invoice documents converted to text, OCR output from your own document scanning pipeline, or programmatically generated invoice data. The text parser handles various formatting styles including tab-delimited tables, fixed-width columns, and free-form invoice layouts with consistent extraction accuracy.

The API extracts over 30 structured fields organized into logical sections. Vendor details include company name, address, phone, email, tax ID (EIN/VAT), and vendor number. Invoice header data covers invoice number, date, due date, PO number, payment terms, currency, subtotal, tax amount, tax rate, shipping charges, total amount, and amount due. Line items are extracted individually with description, item code, quantity, unit of measure, unit price, line total, per-line tax, GL account code, and department assignment. Additional sections include bill-to and ship-to addresses, banking/remittance details, and contract references. All data is returned in a consistent JSON schema with null values for fields that could not be confidently extracted.

The API achieves 96.5% overall field-level accuracy across diverse invoice formats, with higher accuracy for structured fields. Numerical values like amounts, quantities, and tax rates achieve 98.3% accuracy, dates achieve 97.8%, and vendor identification achieves 97.1%. Free-text fields like item descriptions achieve 94.2% accuracy. The system uses a confidence scoring mechanism that returns a score between 0 and 1 for each extracted field. Your application can use these scores to implement validation thresholds — for example, auto-approving high-confidence extractions while routing low-confidence fields for human review. Accuracy improves over time as the ML model encounters more invoice formats from your specific vendor base.

API Codex uses one credit-based subscription that covers all 40 APIs. The Free plan includes 1,000 credits per month, Starter ($29.99/month) includes 50,000 and Pro ($99.99/month) includes 250,000. Each parsed document costs 25 credits, which translates to roughly 40 invoices per month on Free, 2,000 on Starter and 10,000 on Pro. Each request processes a single invoice document, and only successful (2xx) responses consume credits. Custom plans with higher volumes are available on request.

Yes, the Accounts Payable Parser API is designed for seamless ERP integration. The structured JSON output maps directly to the data models of major ERP systems. For SAP, the output aligns with FB60 and MIRO transaction fields including vendor master references, GL account codes, cost centers, and tax codes. Oracle NetSuite integration uses vendor bill API mappings with subsidiary codes and location references. QuickBooks Online integration maps to Bill objects with vendor references, account codes, and class tracking. The API also supports Microsoft Dynamics 365, Sage Intacct, and Xero with appropriate dimension codes and department mappings. Most customers achieve production ERP integration within 1-2 weeks using our standardized output format and platform-specific integration guides.