Fintract

Fintract is a financial document extraction API for developers and finance teams who need structured data from bank statements, invoices, receipts, and financial statements. It is built as a usage-based alternative to cloud OCR platforms, enterprise financial APIs, and in-house OCR-plus-LLM pipelines, for teams that want validated output without a sales process.

A user uploads a PDF to the REST API, and Fintract extracts structured data into a fixed, versioned JSON schema for that document type, with a page number and verbatim quote cited for each critical field. Before the result is returned, the extraction is checked for internal consistency, such as opening balance plus transactions equaling the closing balance, and extractions that fail to reconcile are rejected rather than returned wrong.

Fintract covers bank statements, invoices, receipts, balance sheets, and income and cash-flow statements, each with its own documented schema and built-in checks such as arithmetic validation on invoices and assets-equal-liabilities-plus-equity on balance sheets. Every amount carries its ISO 4217 currency code, and results can be exported as JSON, CSV, XLSX, or, for bank statements, OFX or QBO.

Fintract charges a flat $0.05 per page for every supported document type, with $2.00 of free starter credit (about 40 pages) and no credit card required to begin. Prepaid credit never expires and can be topped up from $25, failed extractions are not billed, and source documents are automatically deleted within 90 days.

Use cases

• Extract transactions from bank statements for reconciliation
• Process invoices with arithmetic-checked line items
• Parse receipts for merchant, totals, and tax
• Verify balance sheets balance correctly