Would love better support for large PDF documents in Agents, especially files with a very high page count.
In some cases, the document is around 1,000 pages and approximately 4 MB. The file itself is not especially large, but the agent can still run into processing limits when trying to read and use the full extracted text in a workflow.
This makes it difficult to use Agents for document heavy use cases such as comparing invoices, statements, or other long reports. We would love to see Agents handle long PDFs more reliably so customers do not need to redesign workflows just to process large documents.
Even in earlier investigation, Support confirmed this type of issue can happen when long PDFs produce very large amounts of extracted text inside an agent step, even when the file is well within the normal upload size limit.
