Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/jhiesey/peercloud

Serverless websites via WebTorrent
https://github.com/jhiesey/peercloud

Last synced: 12 days ago
JSON representation

Serverless websites via WebTorrent

Awesome Lists containing this project

README 

        
# Peercloud



## Decentralized web hosting over WebTorrent



With Peercloud, you can host and view static websites without servers. You download the

site's resources from other regular web browsers just like yours!



For this to work, you need a

browser that supports [WebRTC](https://developer.mozilla.org/en-US/docs/Web/Guide/API/WebRTC) and

[service workers](https://developer.mozilla.org/en-US/docs/Web/API/Service_Worker_API). Currently

this only includes Chrome, although Firefox is adding support for service workers soon.



## How do I use it?



To see a demo, go to the homepage at [peercloud.io](https://peercloud.io)



Simply paste the infohash of a torrent (seeded with WebTorrent) that contains the site you want to

visit and click go.



If you want to host your own site using peercloud, just create a folder containing the files you

want to host, including an index.html file in the root. As long as you use relative paths for all

resources, the site should render exactly the same as if you used any other server to

serve the same static files. For a simple way to seed your torrent, go to

[instant.io](https://instant.io/) and drag your folder onto your browser. The infohash for your

new torrent will appear, which you can then paste into peercloud.



## How does it work?



The peercloud.io site doesn't actually host any content. It just provides the code needed to fetch

individual sites from torrents using [WebTorrent](https://webtorrent.io/). This means that you

have access to the same things whether you go to the official site or any other mirror (including

one you run yourself), so long as the person hosting the mirror is serving the correct code.



Furthermore, when you visit peercloud.io for the first time, your browser will permanently cache

the peercloud.io site itself using a service worker. This means that the site will continue to

function even if you can no longer access the peercloud.io server.



When you supply an infohash and click go, a new tab opens pointing to

`https://peercloud.io/sandbox//index.html`. The browser will then request index.html from

the service worker, which will in turn make peercloud start downloading the torrent and return

the index file. As the site requests more resources, they too will be downloaded from the torrent.



If you request a resource from a relative url, like "myimage.jpg", it will be looked up within your

same torrent. To request a resource from a different torrent, just use a path that starts with the

root directory, like this: `/sandbox//path/within/other/torrent.jpg`



## Limitations



* For security reasons, all pages are sandboxed using

[Content Security Policy](https://developer.mozilla.org/en-US/docs/Web/Security/CSP). Without this,

any torrent could access any resource from any other torrent, which isn't desirable. An unfortunate

side effect is that every resource has an anonymous origin, preventing many web features like

cookies and XHR requests from working properly. I hope to provide a polyfill library to emulate

these features in a safe way at some point in the future.



* Since Chrome doesn't support WebRTC from service workers, you always need a tab open to the main

peercloud.io site when browsing. This means that you can't directly link to a particular torrent,

since that wouldn't open the homepage as well. This can be fixed once Chrome implements WebRTC in

serviceworkers, so if you want to see this fixed, you should

[star this issue on the chromium bug tracker](https://code.google.com/p/chromium/issues/detail?id=302019).



* Service workers only work over https (except on localhost), so if you want to set up a mirror you

MUST get an SSL certificate and configure your server properly.



* This is still very experimental! If you think you can help make it better, please send me a pull

request.



## License



MIT. Copyright (c) John Hiesey.