If you only read feeds, you can skip this article: every reader worth using understands RSS, Atom and JSON Feed, and you will never notice which one a site uses. If you publish a feed, or you are curious, here is how they differ.
RSS 2.0
The oldest and most common. It is XML, it is forgiving, and it is what WordPress, most podcast hosts and many news sites produce. Its specification is short and has been frozen since 2003, which is both its strength and its weakness: some details, such as whether a description contains HTML or plain text, are left vague, and readers have learned to guess.
Atom
Atom was written in the mid-2000s by people who wanted to fix those ambiguities. It became an internet standard, RFC 4287, in December 2005. It requires a unique ID and an update date for each entry, and it says exactly what kind of content each field holds. Blogger, GitHub and many static site generators use it.
JSON Feed
JSON Feed appeared in 2017, with version 1.1 in 2020. It carries the same information in JSON instead of XML, which many developers find easier to produce and parse. Most modern readers support it, but it is far less common than the other two.
Side by side
| RSS 2.0 | Atom 1.0 | JSON Feed 1.1 | |
|---|---|---|---|
| Syntax | XML | XML | JSON |
| Formal standard | Community spec | IETF RFC | Community spec |
| Unique ID per item | Optional | Required | Required |
| Podcasts | The standard | Rare | Possible, rare |
| Reader support | Universal | Universal | Most |
What we would recommend to a publisher
- Use whatever your platform produces. A working feed in any format beats a perfect one you never set up.
- Give every item a stable unique ID. Changing IDs is the main reason readers show old posts as new.
- Publish the full text if you can. Summary-only feeds get read less.
- Announce the feed in your page’s
<head>so readers can find it from your normal address.
And if you want to check that your feed is valid, the W3C Feed Validation Service handles RSS and Atom.
