RSS & Atom Feed Reader + New Item Monitor

Turn any RSS, Atom or JSON Feed into clean JSON rows, or schedule it and get only the new items since the last run. Paste a feed URL or just a website: the feed is found for you.

What does RSS & Atom Feed Reader do?

• Reads RSS 0.9x/1.0/2.0, Atom and JSON Feed (podcasts included) with one parser that tolerates broken XML.
• Auto-discovers feeds from a website or blog URL (<link rel="alternate"> tags, then /feed, /rss.xml, /atom.xml, /index.xml…).
• Outputs flat, typed fields: title, link, ISO-8601 dates in UTC, plain-text summary, categories, author name, image and enclosures (podcast audio).
• Monitor mode remembers what it has already seen per feed and returns only new items. Schedule it hourly or daily and pipe the output to Slack, Sheets, a newsletter or an AI agent.
• Filters by keyword/regex and publication date, so you only pay for items you want.

Who is it for?

• Marketers and founders tracking competitors' blogs, changelogs and press rooms.
• Newsletter and content curators aggregating dozens of sources into one dataset.
• AI / RAG builders who need fresh articles (with full HTML via includeContent) from many publishers.
• n8n, Make and Zapier users who want one reliable node for many feeds instead of one trigger per feed.
• Podcast tools that need episode lists with audio enclosures.

How to use it

1. Add feed URLs or websites to Feed URLs or websites.
2. Pick Read (all current items) or Monitor (only new items).
3. Optional: keywords, "Published after" (24 hours, 7 days, 2026-01-31), max items per feed.
4. Run it, or create a schedule (Monitor mode) and connect an integration.

Input example

{
  "startUrls": ["https://github.blog", "https://feeds.npr.org/1001/rss.xml"],
  "mode": "monitor",
  "monitorName": "competitors",
  "includeKeywords": ["pricing", "launch"],
  "publishedAfter": "7 days"
}

Output example

{
  "id": "https://github.blog/?p=99027",
  "title": "GitHub Copilot app for Beginners: How to build custom workflows with canvases",
  "link": "https://github.blog/ai-and-ml/github-copilot/github-copilot-app-for-beginners-how-to-build-custom-workflows-with-canvases/",
  "published": "2026-09-25T18:00:00+00:00",
  "updated": null,
  "author": "Kayla Cinnamon",
  "summary": "Describe the interface you need in plain English, then let the agent build a live surface…",
  "categories": ["AI & ML", "GitHub Copilot"],
  "enclosures": [],
  "image": null,
  "feedTitle": "The GitHub Blog",
  "feedUrl": "https://github.blog/feed/",
  "site": "https://github.blog",
  "changeType": "new",
  "detectedAt": "2026-09-27T12:00:00+00:00"
}

changeType and detectedAt appear in Monitor mode only. A per-feed summary (format, item count, errors, discovery method) is saved in the key-value store record SUMMARY.

Data fields

Field	Description
id	Stable item id (guid / Atom id / link) used for new-item detection
title, link	Item title (HTML stripped) and absolute URL
published, updated	ISO-8601 UTC timestamps, null when the feed has none
author	Byline name only. E-mail addresses are removed by design
summary	Plain text, cut at summaryChars
contentHtml	Full HTML body, only with includeContent
categories, image, enclosures	Tags, thumbnail, attachments (url, type, length)
feedTitle, feedUrl, site	Where the item came from

How much does it cost?

Pay per event, no subscription:

• $0.001 per item in Read mode (and on a Monitor baseline you ask to output).
• Monitor mode: $0.0005 per feed check plus $0.001 per new item. Quiet feeds cost almost nothing.

Example: 50 competitor blogs checked every day for a month ≈ 1,500 checks ($0.75) plus the new posts they publish (say 300 → $0.30).

Tips

• Use different monitor names for independent watchlists; each name keeps its own memory.
• Feeds only show their latest items. Schedule Monitor mode at least as often as a feed publishes its page size (e.g. a news feed with 10 items and 30 posts a day needs hourly runs).
• Set Feeds per website above 1 to also read category or comment feeds a site advertises.

Known limits

• HTTP only: pages that build their feed link with JavaScript, or feeds behind a login or a bot wall, can't be read.
• robots.txt is respected by default; disallowed feeds are reported in SUMMARY.
• Item history per feed and monitor name is capped at the 5,000 most recent ids.

FAQ

Can it find the feed if I only have the website address? Yes. Most blogs and news sites (WordPress, Ghost, Substack and many more) advertise their feed; the Actor follows that link or tries common paths.

Does it work for podcasts? Yes. Audio files are in enclosures with MIME type and size.

Will I get the same item twice in Monitor mode? No. Item ids are stored after each run, before results are charged, so a stopped run never replays items.

Is it legal? Feeds are published by sites for automated reading. The Actor reads only public feeds, respects robots.txt, and outputs no e-mail addresses. You are responsible for how you reuse the content (copyright still applies to full articles).

Can AI agents use it? Yes, through Apify's MCP server and API; inputs work with only startUrls.