File hosting is an extremely messy business. I don't see any DMCA/abuse reporting mechanism on the site. You need those, and as you scale up, you're also going to need full-time staff whose only job is to wade through uploads and prevent abuse. (No, AI alone can't do that.) Firefox Send had to shut down after becoming a massive malware/phishing distribution center, and that was with full-time staff in that role [1]. Obviously writing a service like this is trivial; keeping it running safely is the hard part.
Considering that your other business is an open, unauthenticated CORS proxy service [2], I have to wonder why you've picked two of the very messiest niches.
Thank you very much for this valuable information! This is currently in a very early stage. DMCA button is the next thing that will be implemented. I learned a lot from running CORSPROXY for a view years now. It is not fully unauthenticated. Only if used for local testing and it has very strict limits. Also all filetypes that are used for illegal streaming are blocked by default. It also has a file size limit. CORSPROXY has like 50+ mechanisms to block illegal content is mainly used to proxy text content like JSON, CSV, etc. The same security mechanisms will be implemented into dlvr.sh very soon or are already implemented.
Building one of these is pretty trivial, especially in the AI era. This one looks like a wrapper for an R2 bucket. The nasty bit is what happens when they start getting abused.
Considering that your other business is an open, unauthenticated CORS proxy service [2], I have to wonder why you've picked two of the very messiest niches.
[1]: https://en.wikipedia.org/wiki/Firefox_Send
[2]: https://corsproxy.io/
https://github.com/pluja/awesome-privacy#file-management-and...
https://en.wikipedia.org/wiki/Category:File_sharing_services
and those are hardly a complete lists