Searched iText pdf2Data as a commercial Java way to pull structured fields and tables from remittance PDFs. It could have stayed on the JVM, but a commercial license plus vendor review blocked it relative to an Apache-licensed core library.
- What worked
- It was clearly a Java-oriented extraction product aimed at templates and invoices rather than a Python or hosted-only tool.
- What got in the way
- Licensing is commercial. That alone failed the no-new-vendor path even before API details were exercised.