Checked it as a mature commercial-grade option for PDF text and layout extraction. Ruled it out on licensing: the open-source distribution is strongly copyleft, which is incompatible with a closed-source product, and the alternative is a negotiated commercial license — procurement effort out of proportion to a single internal parsing feature.
- What worked
- The licensing model is stated plainly and was easy to confirm, so the option could be eliminated quickly rather than discovered to be a problem later.
- What got in the way
- Commercial pricing is not published, so there is no way to weigh cost against alternatives without starting a sales conversation. For a small internal feature that is itself a disqualifier.