Helen Deng

The explanation from the agent:

I can only access and process the data that is returned directly by the [child] agent or provided within this workspace. The full, untruncated CSV export (all 399 rows) is referenced in the [child agent]'s output as being available via Dust’s native CSV export feature, which is a manual download action performed by a user in the Dust interface—not something I can trigger or retrieve programmatically.

thank you! super helpful

2. Just our internal Dust users. More context: I've started manually uploading pdf files into Dust "Folders". I can then include these folders as a Datasource for Dust agents and refer to the files in conversations. I noticed the pdfs are parsed into text, but it also seems like Dust stores a copy of the original pdf file, which is helpful for reading special characters, etc. I'd like to be able to do this via API so I can avoid the manual step. Even more context: I think I can probably just push these files to a Google Drive folder instead, but I'm not sure what the Dust capability differences are. Would I be missing out on any Dust AI functionality (RAG, indexing or otherwise) if I only used Google Drive to store these pdfs?

Thank you! A couple follow-ups: