RSS Feed Monitor – New Items Only, Keyword Alerts

Turn any RSS or Atom feed – or just a website URL – into clean JSON. Filter by keywords, and in monitor mode get only items you have not seen before. Built for news alerts, AI agents and Make / Zapier / n8n automations.

What does RSS Feed Monitor do?

• Reads RSS 2.0, RSS 1.0 (RDF) and Atom and normalizes all of them into one schema
• Auto-discovers feeds from a homepage – paste https://techcrunch.com/ and it finds the feed
• Keyword include/exclude filters (case-insensitive, word-start match: "agent" matches "agents", "ki" does not match "Ukraine")
• Max age filter – e.g. only items from the last 24 hours
• Monitor mode (onlyNew) – remembers items across runs; schedule it and receive only new items
• Clean text – HTML stripped, entities decoded, e-mail addresses removed from summaries
• Broken feeds are reported per URL and not charged

Who is it for?

• Marketers and PR teams tracking brand, competitor or industry news
• Analysts and researchers monitoring blogs, journals or arXiv categories
• Developers and AI agent builders who need a reliable "what's new" tool
• No-code automators sending alerts to Slack, Telegram, e-mail or Notion

Use cases: RSS keyword alerts and news monitoring

• 🔔 Keyword alerts – "AI", "LLM", your brand name → Slack every hour
• 📰 News aggregation – merge dozens of feeds into one deduplicated dataset
• 🤖 AI agent tool – "what's new on these sites since yesterday?"
• 🧪 Research monitoring – new papers from arXiv RSS feeds filtered by topic
• 📈 Competitor tracking – watch competitors' blogs and press pages

How to monitor RSS feeds in 4 steps

1. Click Start – the prefilled input (Hacker News front page + techcrunch.com, max 10 items per feed) runs in a few seconds.
2. Add your feeds or websites and optional keywords.
3. For alerts: set onlyNew: true, choose a monitorId, and add a schedule (e.g. hourly).
4. Connect Slack, e-mail, Google Sheets or your webhook.

Input example

{
  "feedUrls": ["https://hnrss.org/frontpage", "https://techcrunch.com/", "https://rss.arxiv.org/rss/cs.AI"],
  "keywords": ["AI", "agent", "LLM"],
  "excludeKeywords": ["sponsored"],
  "maxAgeHours": 72,
  "maxItemsPerFeed": 50,
  "onlyNew": true,
  "monitorId": "ai-news",
  "includeContent": false
}

Field	Meaning
feedUrls	RSS/Atom feed URLs or website URLs (feed auto-discovery)
keywords	Keep only items whose title or summary contains at least one keyword. Empty = all
excludeKeywords	Drop items containing any of these words
maxAgeHours	Ignore older items (0 = no limit)
maxItemsPerFeed	Upper limit per feed (cost cap)
onlyNew + monitorId	Monitor mode: only items not returned in earlier runs of the same monitor
includeContent	Add the full item content as plain text (if the feed provides it)

RSS to JSON: output example

Real record from a test run on 21 Oct 2026 (keywords KI, agent):

{
  "id": "0560d9c4206f8ceae1f748617ef3379ead708ccd",
  "feedUrl": "https://www.heise.de/rss/heise-atom.xml",
  "feedTitle": "heise online News",
  "feedFormat": "atom",
  "title": "Nvidia will KI-Agenten per Hardware überwachen",
  "link": "https://www.heise.de/hintergrund/Nvidia-will-KI-Agenten-per-Hardware-ueberwachen-11494932.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag",
  "published": "2026-10-21T13:08:00.000Z",
  "summary": "Offene KI-Modelle versprechen Unabhängigkeit. Mit OpenShell und speziellen Prozessoren will Nvidia jedoch die Handlungsräume von Agenten begrenzen.",
  "categories": [],
  "imageUrl": null,
  "matchedKeywords": ["ki", "agent"],
  "fetchedAt": "2026-10-21T19:40:32.260Z"
}

A website URL resolves to its feed, e.g. https://www.tagesschau.de/ → feedUrl: "https://www.tagesschau.de/index~atom.xml".

In monitor mode, a second run right after the first returned 0 items – exactly as intended – and only costs the start fee.

How much does RSS monitoring cost?

Pay-per-event pricing – platform compute is included:

Event	Price
Actor start	$0.005 per run
Feed item returned	$0.001 per item ($1 per 1,000)

Items skipped in monitor mode, filtered-out items and failed feeds cost nothing. The start fee is charged once per run, not per feed.

• Prefilled test run (2 feeds, max 10 items each): ≤ $0.025
• Daily monitor, 20 new items per day: ~$0.75 per month
• Hourly monitor, 20 new items per day: 24 × $0.005 + 20 × $0.001 ≈ $0.14 per day (~$4 per month)

Use maxItemsPerFeed and maxAgeHours as cost caps.

RSS feed API: Python, JavaScript, cURL

Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("mabun/rss-feed-monitor").call(run_input={
    "feedUrls": ["https://hnrss.org/frontpage"], "keywords": ["LLM", "agent"],
    "onlyNew": True, "monitorId": "hn-llm"
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["title"], item["link"])

JavaScript

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('mabun/rss-feed-monitor').call({ feedUrls: ['https://rss.arxiv.org/rss/cs.AI'], maxItemsPerFeed: 50 });
const { items } = await client.dataset(run.defaultDatasetId).listItems();

cURL

curl -X POST "https://api.apify.com/v2/acts/mabun~rss-feed-monitor/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"feedUrls":["https://hnrss.org/frontpage"],"maxItemsPerFeed":10}'

Use with AI agents (MCP)

Give Claude, ChatGPT, Cursor or LangChain agents a "read the latest news" tool via the Apify MCP server: https://mcp.apify.com?tools=mabun/rss-feed-monitor Example prompt: "What did TechCrunch and Hacker News publish about AI agents in the last 24 hours?"

Integrations

• Schedules – hourly/daily monitoring with onlyNew
• Webhooks – call your endpoint when new items arrive
• Make, Zapier, n8n, Slack, Google Sheets, Airtable – via Apify's ready-made integrations

FAQ

What if a site has no feed? Auto-discovery looks for <link rel="alternate"> feed tags. If a site publishes none, the URL is reported as an error (free of charge).

How does monitor mode remember items? Item IDs are stored in a named key-value store in your account, separately per monitorId. On the first run with a new monitor you get the current items; afterwards only new ones.

Can I monitor hundreds of feeds? Yes – add them all to feedUrls. The start fee is charged once per run, not per feed.

Does it return full articles? No – titles, links, summaries and (optionally) the content the feed itself provides. It does not scrape article pages.

Does it work with podcast or YouTube feeds? Yes. Tested on 22 Oct 2026: a podcast RSS feed (NPR Planet Money) returned episode titles, links and summaries; a YouTube channel feed (https://www.youtube.com/feeds/videos.xml?channel_id=…) returned video titles, links and dates, but no descriptions. Audio and video files are not downloaded.

Limitations

• Only content the feed publishes; no full-article scraping
• Some sites block automated requests or publish malformed feeds – these are reported as errors
• Keyword matching works on title and summary (plus content if includeContent is on)
• Feed content remains the publisher's copyright – use it in line with their terms

Related Actors by the same developer

• PDF to Text & RAG Chunks – turn linked PDFs (e.g. arXiv papers) into text for LLMs
• EU Tenders Monitor Germany & Austria (TED) – new public tenders, same monitor mode
• EPO Patent Search & Monitor – new European patents, same monitor mode

Support

Feed not working? Open an issue on the Issues tab with the URL.