Resources
Blog
Notes from building a document extraction product: what accuracy really measures, how to run a monthly receipt routine, and what to check before trusting any OCR vendor.
What invoice OCR accuracy actually means
Vendor accuracy claims are rarely comparable. How field-level accuracy is measured, why confidence scores matter more than averages, and how to test on your own invoices.
June 18, 2026 · 7 min read
- Workflows
A monthly routine for turning receipts into a spreadsheet
Capture, extract, review, archive. A repeatable monthly routine for freelancers and small teams that replaces the quarterly shoebox session.
May 2, 2026 · 6 min read
- Product
Extracting invoice data in Arabic, French and mixed scripts
Right-to-left layouts, Eastern Arabic numerals and bilingual invoices break parsers built for one script. What actually goes wrong, and how it is caught.
March 21, 2026 · 8 min read
- Security
GDPR questions to ask before uploading invoices to an OCR service
Supplier invoices contain personal data. Five questions to put to any document-processing vendor before your first upload, and what a good answer sounds like.
January 29, 2026 · 5 min read
- Accuracy
Why line-item extraction is harder than reading the total
Totals sit in predictable places. Tables do not. Borderless columns, wrapped descriptions and page breaks are what actually separate a parser that works from one that demos well.
November 14, 2025 · 6 min read
- Developers
Choosing an OCR API: a developer's checklist
Latency, idempotency, error envelopes and versioning decide how much of your time an integration consumes. What to check in the documentation before you commit.
September 8, 2025 · 7 min read