RSS & Atom Feed Reader — Auto-Discover Feeds to JSON

Give it feed URLs or plain website URLs. Get back one clean row per item — title, link, ISO 8601 dates, author, categories, plain-text summary and podcast enclosures — ready to export to CSV, JSON or Excel, or pull via API.

What it does

• Reads every common feed format: RSS 2.0, RSS 1.0 / RDF and Atom
• Finds feeds automatically from a website URL: reads the <link rel="alternate"> tags in the page head, then falls back to /feed, /rss, /rss.xml, /atom.xml, /feed.xml and similar paths
• Normalises dates to ISO 8601 (UTC), whatever format the feed uses
• Plain-text summaries (and optional full content): HTML, entities and tracking images are stripped
• Podcast-ready: returns the enclosure URL and MIME type for each episode
• Filters: only items newer than a date or age (2026-01-31 or 7 days), and an optional keyword filter
• Tolerates malformed XML: bare ampersands, control characters and broken items are cleaned up or skipped, and the rest of the feed is still returned
• No browser, no proxies — plain HTTP, so it's fast and cheap

Use cases

• News and competitor monitoring
• Feeding fresh articles into a RAG or LLM pipeline
• Content curation, newsletters and social scheduling
• Podcast episode lists and software release tracking (release Atom feeds work too)
• Building a searchable archive of blog posts

Input

Field	Description
urls	Feed URLs and/or website URLs. Feeds are discovered automatically for websites.
maxItemsPerFeed	Cap per feed. Default 50.
newerThan	Optional. A date (2026-01-31) or relative age (7 days, 12 hours). Items with no date are skipped when set.
keywords	Optional. Keep items whose title, summary, categories or content contain any of these (case-insensitive).
includeContent	Add the full body as plain text, where the feed provides it. Default off.
discoverFeeds	Look for feeds when a URL is a web page. Default on.
maxFeedsPerWebsite	Feeds to read per website when several are advertised. Default 3; comment feeds rank last.

{
  "urls": ["https://feeds.bbci.co.uk/news/rss.xml", "https://techcrunch.com"],
  "maxItemsPerFeed": 20,
  "newerThan": "7 days",
  "keywords": ["AI"]
}

Output

{
  "feedTitle": "TechCrunch",
  "feedUrl": "https://techcrunch.com/feed/",
  "feedType": "rss",
  "title": "Startups don’t need to go viral to win customers",
  "link": "https://techcrunch.com/2026/10/23/startups-dont-need-to-go-viral-to-win-customers/",
  "published": "2026-10-23T23:07:45.000Z",
  "updated": null,
  "author": "Lucas Ropek",
  "categories": ["Startups", "Google"],
  "summary": "At TechCrunch Disrupt last week, two founders and a VC discussed how to not just gain an early following but keep one.",
  "guid": "https://techcrunch.com/?p=3181279",
  "enclosureUrl": null,
  "enclosureType": null
}

With includeContent on, a content field (plain text) is added. A short SUMMARY record in the key-value store lists feeds read and any URLs that failed, with the reason.

Pricing

Pay per item fetched. You only pay for results — feeds that fail or return nothing cost nothing.

Notes

• Reads only public feeds that sites publish for syndication. It does not crawl pages or log in anywhere.
• author is the name exactly as published in the feed. Email-only author fields are ignored, and nothing is looked up or enriched.
• Atom feeds that carry only an updated timestamp (common on software release feeds) return that in updated; published stays null. The date filter uses whichever is present.
• A feed's own item order is kept. Some feeds list only their latest handful of items; that is a limit of the feed, not the Actor.
• This Actor was built and is maintained with AI assistance, with human review. Issues are answered within a few days.