Currently, the platform is limited with uploading CSV and spreadsheet documents only. It would be nice to extend list with jpeg, png and pdf. Those can be automatically converted to PDF. PDF can be parsed with tools like AWS Textract, and fed into LLM of choice, so it can construct csv.

The ask here is not to implement the whole workflow, which I understand might feel pretty cumbersome, but rather somehow let us incorporate it. For instance, let us setup our own file preprocessing webhook, that intakes extended list of file extensions, and outputs csv. That will be a very impressive progress and UX upgrade.