The content is only partially downloaded to the servers, and only on demand. Storage is limited — old, inactive cached data is removed when space is needed for new requests.

There’s also the ability to revisit previously used content via direct links like https://webtor.io/{infohash} — this lets users bookmark a stream or return to it later. However, availability still depends on whether the content is cached or needs to be fetched again.

I actually experimented with making content indexable in the past, but many torrents turned out to be pirated — and eventually triggered DMCA notices. So I chose not to publicly expose anything on the hosted version.

Automation is possible: there’s a public API and a lightweight SDK for embedding content into external websites.