RSS Feed Reader & Monitor: New Items to JSON + Full Text

RSS feed reader and RSS monitor: turn any RSS or Atom feed (or a website that has one) into clean JSON, and get only the new items on every scheduled run.

Monitor news sites, blogs, competitors, release notes, job boards, podcasts or any RSS/Atom feed. Each run returns only the items you haven't seen yet, so you can schedule it every 15 minutes and never pay twice for the same article.

• Pass feeds or plain website URLs. Feeds are auto-discovered from <link rel="alternate"> and common paths such as /feed and /rss.xml.
• Remembers what you've seen across runs, in your own named key-value store. Run separate monitors with different state names. It also uses ETag and Last-Modified, so unchanged feeds cost nothing.
• Full article text as clean Markdown. Navigation, ads and boilerplate are removed, so it's ready for LLM summarisation, RAG or newsletters. The Actor uses the feed's own content when present, otherwise the article page, and respects robots.txt.
• Keyword include and exclude filters (whole-word), a maximum item age, and a per-feed cap.
• Clean, consistent fields across RSS 0.9x/1.0/2.0 and Atom: canonical link (tracking parameters stripped), ISO dates, author, categories, image and enclosures (podcasts).

Typical setups

Use case	Settings
Competitor and industry news alerts	Schedule hourly, add include keywords, connect the Slack or email integration
AI newsletter or digest	fetchFullText: true, then feed the Markdown to your LLM
Podcast and video monitoring	Use the enclosures field for audio and video URLs
Alerts for future posts only	firstRunMode: "baseline" on the first run

Input example

{
  "feedUrls": ["https://github.blog/feed/", "https://simonwillison.net/", "https://feeds.bbci.co.uk/news/technology/rss.xml"],
  "onlyNewItems": true,
  "stateName": "tech-news",
  "fetchFullText": true,
  "includeKeywords": ["AI", "security"],
  "maxAgeDays": 7
}

Output example

{
  "type": "item",
  "feedUrl": "https://github.blog/feed/",
  "feedTitle": "The GitHub Blog",
  "title": "ReviewBench: An open benchmark for AI code review",
  "link": "https://github.blog/ai-and-ml/...",
  "published": "2026-10-05T15:59:40+00:00",
  "author": "...",
  "categories": ["AI & ML"],
  "summary": "...",
  "matchedKeywords": ["AI"],
  "fullTextStatus": "ok",
  "contentMarkdown": "## Why we built ReviewBench\n\n...",
  "wordCount": 1773,
  "firstSeenAt": "2026-10-11T19:05:00+00:00"
}

Feeds that can't be reached produce a free type: "error" row, so you can see broken sources at a glance.

Pricing (pay per event)

Event	Price
New item returned	$0.0005
Full article text extracted (only when it succeeds)	$0.002
Already-seen items, unchanged feeds, errors	free

For example, 20 feeds checked hourly with about 3 new items a day each cost about $0.03 a day, or $0.18 a day with full text. If a run hits your spending limit, items that weren't delivered are kept and returned in the next run.

FAQ

Where is the memory stored? In a named key-value store in your account (rss-monitor-<stateName>). Delete it, or use Reset memory, to start over.

Paywalled or JavaScript-only sites? Full text needs the article in the page HTML. Otherwise you still get the feed's title, summary and link, and fullTextStatus explains why.

Is this allowed? Feeds are published for syndication. The Actor fetches article pages politely, respects robots.txt, and identifies itself. Respect each publisher's terms when you republish content.

---

Built by PewterDesk (pewterdesk.pages.dev). More tools: apify.com/pewterdesk.