My work processes PDFs from government sites, filling them out for end users automatically. Would be cool if the API could do this somehow. Parse fields, and let you put text into them.