In Dust, a 'feed' is a catch-all term for 'something that watches something external and returns a stream of useful messages' (not a very precise definition ...).
The repo contains a small (hopefully growing) library of Actors (primarily Dust pipeline stages where they find most use) for creating feeds. Currently supported are:
- RSS feeds (RSS, Atom etc). Periodically visit an RSS feed and return the latest published documents (or their meta info).
- Web Spider Actor for sites (not the whole internet). Crawl a web site, filter onsite links etc etc and send HTMLDocumentMsgs on for further processing.
- Send queries to search engine aggregator SearxNG and send on responses.
Option 3 probably requires that a local instance of SearxNG be installed unless you can find a public one that reliably supports returning JSON.
Dust-feeds uses the dust-core, dust-http and dust-html libraries, so a prior familiarity with those is probably a good idea.