Seven document types. Seven templates.

An invoice is not a statement, and a packing slip is not a receipt. Cribble classifies every document first, then extracts the fields that matter for that type: line items and bank details for invoices, balances for statements, quantities for packing slips.

Get started
Different kinds of financial documents arranged across a bookkeeper's desk

Classification before extraction

Invoice, receipt, statement, remittance advice, purchase order, packing slip or credit memo: Cribble decides what it's reading before it reads it, so the right fields are populated correctly. Every. Time.

Handwriting included

Typed, printed or handwritten; if a person can read it, Cribble casn too. Confidence scores are consistent with legibility.

Three languages, three currencies at launch

English, German and French documents in USD, EUR and GBP, with language and currency detected and normalized automatically. More on the way.

Why type-aware beats one-size-fits-all

Generic Optical Character Recognition (OCR) templates flatten every document into the same handful of fields, which is why they mangle statements and lose line items. Cribble maintains a separate extraction model per document type, focused in on the fields bookkeepers actually enter manually.

The result: opening and closing balances from statements, PO numbers and requested quantities from purchase orders, credited amounts from credit memos; each in its right place, each individually confidence-scored.

And when the standard fields aren't enough, define your own. Business workspaces can add custom fields — a cost center, a project code, a contract number — with a plain-language note on what to look for. Cribble reads them from every new document alongside the built-in fields, confidence scores, inline correction, exports and webhooks included.