3 ms·
Probably the best use case for onion routing and IPFS. Each X URL gets a content addressable IPFS URL, served via an onion site that should be hard to take down
by toomuchtodo 2mo ago
Probably the best use case for onion routing and IPFS. Each X URL gets a content addressable IPFS URL, served via an onion site that should be hard to take down.
Browser->Tor Onion site->IPFS as storage system pull through cache->X
- flotzam 2mo agoOr just a .onion for the frontend and code forge + another server at some clearnet domain proxying it (until taken down and replaced with the new one) would probably work fine. The hidden service can even terminate TLS with a valid cert for the clearnet domain.
- toomuchtodo 2mo agoTrue, I suggested IPFS as an hidden content addressable, globally distributed caching proxy and persistent object store. Run it from outside the reach of the US legal framework, like private/shadow libraries/archives (as you could consider it an archive of X, with public access when available). Another idea I recently had was extending Karakeep (https://github.com/karakeep-app/ https://github.com/karakeep-app/), which has robust archiving capabilities today, to enable announcing archived URLs via a DHT and serving the archive bundle as a torrent (similar to how the Internet Archive creates a torrent file for each item). With some work, it would be the equivalent of a distributed Wayback Machine, with each instance able to act as both scraper and server. With a browser extension that could query this DHT, you could retrieve an archived version of any URL that has been archived by cherry picking from the swarm (each URL becomes a magnet torrent URL, very similar to IPFS, but with less bootstrapping required imho). Distributed web adjacent (https://getdweb.net/ https://getdweb.net/).