RSS, Atom & JSON Feed Monitor - Only New Items

Check many RSS, Atom and JSON Feed URLs in one run and get one clean dataset with the same fields for every feed. Schedule it with Only new items turned on, and each run returns (and charges for) only the posts that were not delivered before.

What it does

• Reads standard feeds. RSS 0.9x / 1.0 / 2.0, Atom and JSON Feed. You paste the feed URLs; the Actor does not search websites for feeds and does not read normal web pages.
• One consistent output. Every item has the same fields, whatever the feed format: title, link, published (ISO date, UTC), summary (plain text, HTML removed, up to 1,000 characters), author, categories, enclosures (for example podcast audio files), feedTitle, feedUrl and a stable itemKey. Optionally contentText: the full text the feed provides, HTML removed, up to 20,000 characters.
• Only new items (optional). Remembers which items it already delivered, in a named key-value store in your own Apify account, so scheduled runs return only new items. It also sends standard "has this changed?" headers (ETag / Last-Modified), so unchanged feeds are cheap to check.
• Two simple optional filters:
  • Published in the last N hours. Drops items older than N hours. Items with no date in the feed are kept.
  • Keywords (any). A plain text match. An item is kept if its title, summary or categories contain any of your words, ignoring upper/lower case. It matches parts of words too, so art also matches "start". There is no smart or AI matching.
• Clear error rows. A feed that can't be read gets its own row with an error code (for example FEED_HTTP_404, NOT_A_FEED, INVALID_URL, ROBOTS_DISALLOWED, SITE_UNREACHABLE), and the rest of the run continues.
• Polite to websites:
  • Follows robots.txt by default. If a site's robots.txt can't be reached at all, that feed is skipped.
  • At most about 1 request per second to the same site.
  • Retries temporary errors (HTTP 429 / 5xx, network errors) up to 3 times.
  • Ignores feeds larger than 10 MB.
  • Only public http(s) addresses are accepted.

How to use it

1. Paste one or more feed URLs into Feed URLs. Sample input (these are example data sources, not endorsements): https://blog.python.org/rss.xml, https://xkcd.com/atom.xml.
2. Click Start. Results appear in the Output tab, ready to export as JSON, CSV, Excel and other formats Apify supports.
3. To monitor feeds, turn on Only new items since the last run, give the monitor its own State store name, and create an Apify Schedule (for example hourly or daily).

Input

Field	Default	What it does
feedUrls	(required)	Feed URLs. Duplicates are removed.
onlyNewItems	false	Return only items not delivered by an earlier run with the same state store.
stateStoreName	rss-monitor-state	Name of the store that remembers delivered items. Use a different name for each separate monitor.
sinceHours	empty	Keep only items published in the last N hours (1–8,760).
keywords	empty	Keep only items whose title, summary or categories contain any of these words (plain text match).
includeContent	false	Add contentText (full feed text, up to 20,000 characters) when the feed provides it.
maxItemsPerFeed	50	Read at most the first N entries of each feed (feeds usually list the newest first).
maxTotalItems	1,000	Stop after delivering this many items in a run.
maxFeeds	50	Check at most this many feeds per run; extra feeds are skipped.
respectRobotsTxt	true	Skip feeds that the site's robots.txt does not allow.

Output example (real output from a test run)

{
  "feedUrl": "https://blog.python.org/rss.xml",
  "feedTitle": "Python Insider",
  "title": "Python 3.15.0 (final) is here!",
  "link": "https://blog.python.org/2026/10/python-3150-final-is-here/",
  "published": "2026-10-09T00:00:00Z",
  "categories": ["releases"],
  "summary": "A brand new Python here for you to enjoy!",
  "enclosures": [],
  "itemKey": "…",
  "fetchedAt": "2026-10-11T05:45:58Z",
  "error": null
}

Error row example: {"feedUrl": "not a url", "error": {"code": "INVALID_URL", "message": "Not an http(s) URL: 'not a url'"}}

Pricing (pay per event)

Event	Price	When it is charged
feed-item	$0.001 ($1 per 1,000 items)	Each item delivered to your dataset
feed-checked	$0.0002	Each feed fetched successfully, also when there is nothing new
apify-actor-start	$0.00005	Once per run at up to 1 GB memory (Apify's standard start event; one per GB of memory)

• Free of charge: feeds that fail, are unreachable, are blocked by robots.txt, or are invalid. Error rows are never charged.
• Measured example: a test run with 2 feeds and 6 items cost $0.00645 (1 start + 2 feeds + 6 items) and took about 4 seconds.
• Monitoring example (calculated): checking 5 feeds every hour with nothing new costs $0.00105 per run, about $0.76 per 30 days. Each new item adds $0.001.
• Spending caps: set Maximum cost per run in the run options (at least $0.01), and/or maxTotalItems. The run stops cleanly when a limit is reached.
• No separate platform-usage charge: you pay only the events above.

Good to know

• Only new items needs a store. It works when the same State store name is used run after run. A new name starts fresh, and the first run returns the current items.
• Without Only new items, every run returns the current feed items again and charges for them. Turn it on for schedules.
• How long items are remembered. Up to the last 2,000 delivered items are remembered per feed.
• Breakage warning (only with Only new items): if a feed that worked before suddenly returns a web page or something unreadable, the run log shows an ALERT feed_possibly_broken line, and the feed gets an error row. If it suddenly returns zero items, the run log shows the same alert line, with no error row.
• Run summary. Each run saves a RUN_STATS record (counts, errors, charges) in its key-value store.
• Permissions: runs with limited permissions. It only uses its own run storage and the named store it creates for Only new items.
• Content rights: feeds are published for syndication, but the content belongs to its publishers. Respect their terms and copyright if you republish it. The Actor only reads public feed URLs you provide. It does not log in or bypass paywalls.

Support

Found a problem or need a feature? Open an issue in the Issues tab.