[unstatusthequo]

The idea is nice, but uploading documents is a security concern. I get you “use industry standards” to protect data, but that’s vague and any actual real privacy and security people will brush it off. How do we know what’s going on behind the scenes?

So enterprise use will be extremely limited I think. This would be better as a software that runs locally to extract text, or in customer’s own controlled clouds. Maybe use Apache Tika for that, with a nice frontend. Then, make calls so the OpenAI API only using the extracted text. The same reason I won’t upload important documents to those free “file conversion for free!” websites applies here. Just too risky for the reward. But, fix that requirement and you have something nice I think.