I’ve been looking for an easy OCR solution considering I have about 10,000 one-page documents a month (invoices). For comparison, Amazon textract is ~$0.05/pg for key-value pairs, but it involves more programming to set up.
At this point in time the price is per generated document - irrespectively of how complicated the operation is.
Because you can combine operations in one http call, you're incentivised to do that as opposed to perform separate calls which increase the possibility of errors and cost for all sides.
Happily taking feedback though - your comment around hard-cap is definitely sound, for example.
As part of an integration for a few customers we do relatively simple PDF operations. Combining a few TIFFs into a PDF, or splitting a PDF into separate PDF per page for example. Think invoices vs orders. Can be a royal PITA though due to edge cases in terms of whatever generated the TIFFs or PDFs.
However each of our customers have 1-10k documents per day, and we're in Scandinavia so small fries in terms of volume.
edit: OCR and table extraction is something that might be very interesting depending on quality, but again even 10k/mo seems low for most of our larger customers.
Source: I work at PDF Blocks
Quite expensive though. When I use an API, I usually assume (1) there will be some significant base volume, and (2) this volume has no upper bound, depending on my users behavior. For ~750 € you can only process 1k documents during the month... hard-capped? The price schemes seem to target entreprise but entreprises usually have bigger volumes than that. (But maybe I confuse API calls and document processing with your product?)
But it’s nice you released SDK in several common languages.
Good luck!