Generated feeds and syndication

(Almost) anything you see in the headlines buffer can also be accessed and syndicated further via a generated feed (in Atom format) by clicking the little feed icon at the right of the currently loaded feed title; in version:1.12 next to the feed list near the top.

older versions	1.12 and newer

You can access labels, categories, tags, search results as RSS feeds. Since version:1.5.0 you can also limit the articles which go into generated feeds, e.g. you can have a generated feed which includes published articles from category X.

Known issues

Subscribing to your own feed of search results will cause duplicate articles on each and every feed update. See this forum thread for details: https://discourse.tt-rss.org/t/behaviour-of-search-feeds/122/2

Data protection

Generated feeds are protected using random unique keys. Key can be regenerated at any time.

Anatomy of a generated feed URL

Consider an example:

http://x.y.z/tt-rss/backend.php?op=rss&id=61&is_cat=1&view-mode=adaptive

id (integer) - requested feed ID
is_cat (boolean) - whether the feed is a category
view-mode (string) - see below
key (string) - automatically generated access key (since version:1.5.0)

Optional parameters:

login, pass - see above
format - since version:1.6.0 specifies output format, possible values: atom, json
limit - amount of articles to output, default: 30
offset - start output while skipping this amount of articles, default: 0 (since version:1.6.0)
order - override default headlines order (since version:1.8)
ts - output articles newer than timestamp in strtotime accepted format (since version:1.12) i.e. stuff like ts=1%20month%20ago

Special feed IDs:

1 Starred articles
2 Published articles
3 Fresh articles
4 All articles
0 - Archived articles

Feed IDs < ~~10 are considered Labels. \ Special category IDs (is_cat=1): \ * 0~~ Uncategorized\ * 1 Special category (includes Starred, Published, etc.)\ * 2 Labels category (includes your labels)

View mode values:

Note: It’s probably not a very good idea to use Adaptive view mode for generated feeds.

adaptive - shows unread articles only when they are unread articles, shows everything otherwise
marked (this means starred), has_note, published, unread, unread_first - should be self explanatory

Actual output may differ between modes for several special feeds for usability reasons, e.g. recently read feed ignores unread specifier because unread articles are never part of the feed).

Order values:

default - depends on the feed: either import batch date or (for published and starred feeds) last_published and last_marked
title - sort by title
date_reverse - reverse sort by batch date
feed_dates - sort by feed-provided article dates