RSS Feed Reader: bulk RSS, Atom & JSON feeds to a clean dataset

Read hundreds of RSS, Atom, RDF and JSON feeds in one run and get one clean, normalized row per article, post, episode or video. Don't know the feed URL? Just paste the website: the Actor finds the feed automatically (<link rel="alternate"> tags, then common paths like /feed, /rss.xml, /atom.xml).

Turn on only-new monitoring and schedule it, and every run returns just the items published since the last run, ready to pipe into Slack, email, Google Sheets, a webhook, Zapier, Make or an LLM.

Why this feed reader

• All formats, one schema: RSS 0.9x/2.0, RSS 1.0 (RDF), Atom 1.0 and JSON Feed 1.x, including podcast enclosures, Media RSS images and YouTube channel feeds.
• Feed autodiscovery from any website or blog URL.
• Only-new items between runs (per-feed memory, separate monitors via stateKey).
• Filters: keywords, last N days or a fixed date, per-feed and total limits.
• Full content (HTML and optional plain text for LLM/RAG pipelines), images and enclosures.
• Cheap and fair: $0.30 per 1,000 items. Feeds that fail and invalid inputs are free and still reported.
• Light (256 MB, plain HTTP, no browser), so runs are fast and cost little.

Input example

{
  "feedUrls": ["https://hnrss.org/frontpage", "https://blog.cloudflare.com", "https://www.theverge.com/rss/index.xml"],
  "maxItemsPerFeed": 20,
  "keywords": ["AI", "security"],
  "onlyNewItems": true,
  "stateKey": "tech-news"
}

Output example

{
  "feedTitle": "The Verge",
  "feedUrl": "https://www.theverge.com/rss/index.xml",
  "siteUrl": "https://www.theverge.com/",
  "feedFormat": "atom",
  "itemId": "3f1c0a8e9b2d4c71",
  "title": "Windows 11 tries new shortcuts for Xbox gaming mode",
  "url": "https://www.theverge.com/...",
  "guid": "https://www.theverge.com/...",
  "published": "2026-10-31T15:12:19.000Z",
  "updated": "2026-10-31T15:12:19.000Z",
  "author": "Terrence O'Brien",
  "categories": ["Gaming", "Microsoft", "Tech"],
  "summary": "Plain-text summary…",
  "contentHtml": "<p>Full article HTML when the feed includes it…</p>",
  "imageUrl": "https://platform.theverge.com/wp-content/uploads/...jpg",
  "enclosures": [],
  "commentsUrl": null,
  "inputUrl": "https://www.theverge.com/rss/index.xml",
  "fetchedAt": "2026-10-31T16:50:00.000Z"
}

Rows for inputs that could not be read contain inputUrl, feedUrl and error and are not charged. A per-feed report (items found, returned, skipped as old/seen/keyword) is saved as REPORT in the run's key-value store.

Pricing

Pay per event: $0.0003 per item returned ($0.30 per 1,000) plus a tiny Actor-start fee. Set Maximum cost per run and the Actor stops cleanly at your limit.

Use cases

• News and brand monitoring: keyword alerts across many publishers.
• Content aggregation for newsletters, dashboards and apps.
• Feeding LLM/RAG pipelines with fresh articles (use includeContentText).
• Podcast episode and YouTube upload tracking (https://www.youtube.com/feeds/videos.xml?channel_id=…).
• Competitor blog and changelog tracking.

Monitoring setup (only new items)

1. Turn on Only new items and give the monitor a state name.
2. Create an Apify Schedule (e.g. hourly).
3. Add an integration (Slack, email, webhook, Google Sheets). The first run returns the current items; later runs return only new ones.

FAQ

Does it scrape websites? No. It reads the public feeds that sites publish for exactly this purpose. Autodiscovery reads only the page's <head> links and a few standard feed paths.

Why did a site return an error? Some sites have no feed, or block automated requests. Error rows are free.

Items without dates? They are kept even when you use date filters.