RSS & Atom Feed New-Item Monitor
Under maintenancePricing
from $2.00 / 1,000 new item detecteds
RSS & Atom Feed New-Item Monitor
Under maintenanceWatch any RSS, Atom, RDF, or JSON feed for genuinely new entries. Built-in cross-run dedup means scheduled re-runs push nothing and cost ~nothing when nothing new appears. Pay only per new item detected.
Pricing
from $2.00 / 1,000 new item detecteds
Rating
0.0
(0)
Developer
Delta Tools
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
19 days ago
Last modified
Categories
Security
Limited permissionsShare
Watch any RSS, Atom, RDF, or JSON feed for genuinely new entries — with built-in cross-run deduplication. Schedule it hourly or daily and get alerted only when a feed publishes something new. Quiet runs push nothing and cost ~nothing.
Why this instead of a feed reader
One-shot feed readers re-bill you for every entry on every scheduled run, and
leave the "what's actually new?" diffing to you. This Actor keeps a snapshot
of every entry it has already seen (per stateKey), so:
- Only genuinely new entries are ever output — repeat runs with nothing new return an empty dataset.
- You pay per new item, not per entry scanned. A quiet hourly run costs only the Actor start ($0.001).
- No API key, no login, no proxy. Plain HTTP fetches of public feeds.
Inputs
| Field | Description |
|---|---|
feeds | One entry per feed URL (max 50). Each has url (required — RSS 2.0, Atom, RDF, or JSON Feed), optional name (friendly label; defaults to the feed's own title), and maxItems (5–500, default 50) for how many of the newest entries to compare. |
stateKey | Namespace for the saved snapshot. Use a different key per schedule so monitors never share baselines. The prefilled default key rotates daily so the free trial always shows recent entries. |
reportInitialAsNew | On the very first run there is no baseline, so every entry is reported as new. Turn off to silently build the baseline instead. |
httpTimeoutSeconds | Per-request timeout (default 45). |
Feeds publish newest-first, and entries are diffed by identity: guid/id
first, then the entry link, then a title+date hash. An edit to an already-seen
entry does not re-report it.
Outputs
One dataset row per new entry: feedUrl, feedName, entryId, title,
link, published, updated, author, summary (first 500 chars),
categories, detectedAt.
Pricing
Pay-per-event. You are billed per genuinely new entry (new-item
event). Entries that were already seen are never billed. A run where nothing
is new bills only the Actor start.
Honest limitations
- Only public feeds reachable over plain HTTP(S) — no authenticated or paywalled feeds.
- Identity falls back to title+date hash when a feed provides neither guid nor link; two entries with identical titles and dates on such a feed would dedup together.
- Feeds larger than 5 MB are truncated before parsing.
- A feed that renumbers its guids (rare, usually after a CMS migration) will
look like all-new entries on the next run — change
stateKeyor accept one noisy run after such a migration. - Very high-frequency feeds (hundreds of entries/hour) can outrun an hourly
schedule: raise
maxItemsor schedule more often.
