RSS

Many websites publish new posts continuously. Anyone who doesn’t want to miss anything would have to visit each site individually and check for updates. RSS solves this problem the other way around: the site places an additional file that neatly lists its latest posts. For each post it states the title, date, a short summary, and the link to the full text. A program on your device fetches this file regularly and shows you everything new in a single list. The abbreviation stands for “Really Simple Syndication.”

News without an algorithm in between

The most important difference from social networks lies in the selection process. On Instagram or TikTok, a recommendation system decides which posts you see and in what order. With RSS, you alone decide which sources you subscribe to. Everything these sources publish lands in your list, usually simply sorted by time. There is no one sorting, filtering, or inserting ads in between.

On top of that comes independence from individual companies. RSS is an open standard that no one owns and for which you don’t need an account. If your reader app is discontinued, you switch to another one and take your subscriptions along as a file. That is precisely why the format has endured since the early 2000s, even though major providers dropped it along the way. Google shut down its popular Reader in 2013, yet the format itself survived without issue.

For the AI industry, RSS is interesting for another reason as well. Anyone who wants to process current news automatically needs it in machine-readable form. A program has to laboriously take apart a normal webpage, because text, menus, and ads are all mixed together. An RSS feed, by contrast, already delivers cleanly separated fields.

What’s inside such a feed file

A feed is a text file with a fixed structure, usually in the XML format. XML means that every piece of information is framed by markers, similar to labeled drawers. This lets a program reliably recognize which part is the title and which is the date. At the very top are details about the source itself, followed by the individual posts as entries.

Retrieval works on a pull principle. Your reader app queries the feed’s address at fixed intervals, say every hour. It compares the list with what it saw last time and marks everything new as unread. The website itself sends you nothing and basically doesn’t know that you’re reading it. This is the opposite of a push notification, where the provider actively sends something to your device.

Alongside RSS there exists the very similar format Atom, which is technically defined somewhat more cleanly. In everyday use the difference hardly matters, since practically every reader app understands both. Colloquially, people call everything “RSS feed” anyway.

Podcasts, newsreaders, and data pipelines

Today RSS is most widely used in podcasts, without most listeners knowing it. At its core, every podcast is an RSS feed in which audio files are linked instead of articles. Your podcast app checks this feed and downloads new episodes. Because the format is open, the same podcast can appear simultaneously in Spotify, Apple Podcasts, and any other app.

To read news, people use programs like Feedly, NetNewsWire, or the open-source Miniflux. You add the addresses of the sites that interest you there, and get a combined news stream. Many editorial offices, government agencies, and blogs still offer their feed, often inconspicuously linked in the footer of the page.

In the tech industry, RSS also shows up as a building block in automated workflows. A program reads feeds from several news sites, sends the texts to a language model, and has them summarized or sorted by topic. Stock alerts, train announcements, or software updates are also distributed this way. A common misconception is that RSS is dead: it has merely become invisible and keeps working in the background.