I wanted to use IPFS because it's API was dead simple to implement as a ShareX shortcut that lived in my context menu and browser extension and even worked through my Headscale VPN and would result in an easy HTTP URL that I could share (I use a VPS to reverse_proxy into my VPN). For small files, it usually works. But for big files, say 50mb+, it started to be wonky. This 2GB file threw it off - which it really should not.

I am on the same network as the node, writing to a storage device absolutely capable of handling it and streaming off that through a connection that actually unterutilizes that storage devices by a lot, and it still fails to stream the file.

My friends aren't as tech-savy as me, so if I can give them just a link, that's perfect. They click it and get a file and I don't have to fight with Discord's stupid upload limitations (let alone buy Nitro...ew...). But this isn't the first time IPFS has failed to serve this need, which disappoints me as it's primary and main purpose is to be a file host - be it as a single-node use as mine was, or in a distributed way where each node should be able to reach every file in the DHT that it can find peers for. So, it is a file solution first and foremost. And it just can't deliver at bigger scale (in terms of filesize), so it failed it's core function.

This probably comes across as ranty... sorry. ^^;

But as far as I know, both me and my friend need to run OnionShare, and I can't just shut my PC down mid transfer, I need to keep it on. His internet is slow, so the idea was to drop stuff into my IPFS node, send him a link, and let garbage collection take care of the rest and I could go to sleep while he figures things out on his end. But, if downloads fail like that, then I can't use this solution anymore, which sucks. :/