Media Analysis via MCP

Media in my use case really refers to PDFs, but using the broader term in case it could have larger impact. I subscribe to external AI tools (Claude) and have built up an extensive PDF library in Capacities since I started using the program. With the limited AI budget included in Capacities Pro, I'd love to be able to use my existing limit outside of the program to analyze these PDFs and then inject the analysis back into Capacities. I could do this just by pointing my agent at the PDF object or by giving it the full PDF directly and telling it which object should take the analysis, whichever has a lower implementation burden.