Python kit · Technical beta · $9

Invoice PDFs into review-ready rows.

Extract candidate invoice data from text-based digital PDFs to CSV and JSON, then compare it with the original before anything enters your books.

Local Python workflowBring your own model endpointHuman approval required

Checkout and file delivery are handled by Gumroad. Your model provider may charge separately.

Invoice data flowing into a reviewable table
What you receive

A small, inspectable starting point.

Standalone Python kit

Run a text-PDF extraction workflow with your own compatible chat-completions endpoint. No hosted account with us.

Review-only outputs

CSV and JSON candidates with arithmetic checks. The tool does not approve invoices or write to your accounting system.

Setup, tests and examples

Offline tests and five clearly synthetic sample invoices let you inspect the mechanics before using private documents.

Inspect before purchase

See where the human review happens.

Synthetic invoice review demo
Internal demonstration using synthetic data. This is not a customer accuracy benchmark.

Provide a text PDF

Use a digital invoice with selectable text; scans need OCR outside this kit.

Inspect candidate fields

The configured provider proposes rows and the script checks basic arithmetic.

Compare and export

Review against the original invoice, then use CSV/JSON only when you accept the result.

Important boundaries
  • No OCR, n8n workflow, direct spreadsheet write or automatic invoice approval.
  • Invoice text is sent to the model provider you configure; check its charges, privacy and retention terms.
  • Compatibility with a live commercial provider and customer accuracy have not been independently benchmarked.

Need a reviewable starting point?

Buy the beta, inspect the samples, and decide whether it fits your workflow.

Get it on Gumroad · $9 ↗