Serverless Sync in Web Apps Using Bit Torrent

Neat use of web torrent. Now if only we could do it with live data instead of persistent data, have built in encryption, and anonymize the peers we'd have the perfect decentralized internet :-)

Now someone could make a Chrome plugin for YouTube, which seamlessly seeds the watched videos and, when it goes offline, loads from Bittorrent.

Why not use IPFS, which is specifically designed for this sort of usage and will be more resilient?

Very cool. I imagine when bittorrent first starting how cool this would be to do if we had something like webtorrent<p>Two problems:<p>1) WebTorrent isn&#x27;t very persistent. It requires people keep the tab open and I don&#x27;t think the current WebTorrent to existing torrent clients federation really solves anything interesting. Unless you&#x27;ve got high volumes of traffic it doesn&#x27;t help much.<p>2) Bandwidth and storage got so cheap. S3 et al is really pricey but you can get insane deals on colo and bare mental boxes these days. Ok, you&#x27;re not going to get s3 ease of use and reliability, but webtorrent is going to be way worse on both those points.

So I only kind of understand magnet links. But it seems like I could create a Chrome plugin that took every page I visited and turned the assets into magnet links and started request&#x2F;seeding them on pageload. Is that analogous to what they&#x27;re doing?

This is really cool. Wonder how hard it would be to develop double-ratcheted multi-user chat or a distributed game using WebTorrent.

Anyone have suggestions on how to block sites like this from downloading without permissions?<p>edit: So blocking XHR requests seems to do the trick.

The app is quite interesting and I was amazed until I realized the play button doesn&#x27;t display the circular progress indicator on time. It has an ugly lag. This is why web apps are not quite there yet. And yet it is amazing they can record, store and sync audio like that. Almost like real apps.

Maybe this a bit tangential, but I was just thinking lately how it would be nice to use some decentralized communication backend to word processing software to enable shared collaborative editing of documents. Sort of like Google Docs, but without a centralized server.

If there wasn&#x27;t a &#x27;cloud&#x27; in the picture how seeding is done I&#x27;d even believe it&#x27;s serverless. Sorry for being a cynic, but what if there is no tracker available?

I was surprised to find out that the torrent fetching part of the prototype didn&#x27;t happen in the service worker . Which is obviously the most cool place for it.

This sounds really cool, but it also sounds like a really easy way for malware (delivered, say, in a drive-by malicious ad) to exfiltrate data from a browser.

It&#x27;s great to share data, however, how can a torrent solution support services like web API?