RSS / Atom Feed Normalizer

Fetch and normalize public RSS 2.0, Atom 1.0, and RDF/RSS 1.0 feeds into consistent structured JSON. Auto-detects feed format and returns a unified schema regardless of source. Atom article links resolve relative references and inherited xml:base against the final feed URL, so moved feeds remain usable in content pipelines.

Use cases

• Content aggregation - merge feeds from multiple sources into one consistent schema
• Feed monitoring - track updates across many feeds with a single API call
• Marketing automation - pull blog and news feeds into scheduling pipelines
• News aggregation - normalize heterogeneous feeds before loading into a data warehouse
• Brand monitoring - track RSS feeds for mentions and updates

Input

Field	Type	Required	Default	Description
feedUrls	array<string>	yes	-	Public RSS/Atom/RDF feed URLs (max 50 per run)
maxItemsPerFeed	integer	no	20	Max items to normalize per feed (1-100)
timeoutSeconds	integer	no	10	Whole-feed fetch deadline, including DNS and redirects (3-30 seconds)

Example input

{
  "feedUrls": [
    "https://hnrss.org/frontpage",
    "https://blog.cloudflare.com/rss/"
  ],
  "maxItemsPerFeed": 20,
  "timeoutSeconds": 10
}

Output

Each feed produces one dataset item:

Field	Type	Description
inputUrl	string	The feed URL provided
feedUrl	string	Normalized requested feed URL, before redirects
ok	boolean	Whether fetch and parse succeeded
status	integer	HTTP status code
format	string	Detected format: rss2, atom, rdf, or unknown
finalUrl	string	Final URL after redirects
contentType	string	HTTP content-type header
checkedAt	string	ISO timestamp
channel	object	Feed metadata: title, link, description, itemCount, itemsReturned
items	array	Normalized feed items (see below)
error	string	Error message on failure

Normalized item fields

Every item has the same fields regardless of feed format:

Field	Type	Description
title	string	Item title
link	string or null	Item URL; for Atom, an absolute HTTP/HTTPS alternate link, or null when none is usable
guid	string or null	Unique identifier; RSS GUID text and Atom IDs remain strings, including numeric-looking IDs with leading zeroes. Missing/empty identifiers return null
description	string	Summary or content text
pubDate	string	Publication date in ISO 8601 format
author	string	Author name or email
categories	array	Category or tag strings

Example output

{
  "inputUrl": "https://hnrss.org/frontpage",
  "feedUrl": "https://hnrss.org/frontpage",
  "ok": true,
  "status": 200,
  "format": "rss2",
  "channel": {
    "title": "Hacker News: Front Page",
    "link": "https://news.ycombinator.com/",
    "description": "Hacker News RSS",
    "itemCount": 30,
    "itemsReturned": 20
  },
  "items": [
    {
      "title": "Show HN: A new tool",
      "link": "https://news.ycombinator.com/item?id=12345",
      "guid": "https://news.ycombinator.com/item?id=12345",
      "description": "Description text...",
      "pubDate": "2026-07-06T12:00:00.000Z",
      "author": null,
      "categories": []
    }
  ]
}

Security

• Only public HTTP/HTTPS feed URLs are accepted
• SSRF protection: localhost, private IPv4/IPv6 ranges, and DNS-resolving-to-private IPs are blocked
• URLs with embedded credentials (user:pass@host) are rejected
• Each redirect is revalidated before fetching (maximum five); sockets use the validated DNS address while retaining the original hostname for HTTP and TLS
• Feed bodies capped at 2,000,000 bytes while streaming; oversized feeds fail rather than returning truncated XML
• Requests ask for uncompressed responses; unsolicited compressed bodies are rejected
• The whole-feed deadline includes DNS, all redirect hops and response reading
• Article links, xml:base targets and embedded content are never fetched. Feed text/HTML is untrusted and must be sanitized before rendering
• No browser automation, no proxies, no login credentials

Pricing

Pay-per-event pricing:

Event	Price
Actor start	$0.005 per GB of Actor memory, minimum one event
Feed result (including error results)	$0.005

At the default 512 MB memory, a run returning 10 feed results costs approximately $0.055. Apify's displayed pricing is authoritative; results are charged per feed dataset row, not per article.

FAQ

What feed formats are supported? RSS 2.0, Atom 1.0, and RDF/RSS 1.0. The actor auto-detects the format and returns it in the format field.

What if a feed is down or returns an error? The actor returns an item with ok: false and the error details. Other feeds in the same run continue normally.

Is there a limit on feed size? Feed bodies are capped at 2,000,000 bytes per fetch. XML nesting and entity-expansion limits also apply, so a feed within the byte limit can still be rejected for parser safety.

Which Atom URL is returned? channel.link and each item's link use rel="alternate" (including omitted rel), preferring HTML/XHTML representations, then source order for ties. Feed, entry and link-level xml:base values are inherited. Without an explicit base, the final feed URL after redirects is used. Self, enclosure and related links are not substituted for article links. Missing, malformed, credentialed or non-HTTP/HTTPS alternatives return null. Article destinations are metadata, not independently DNS-validated or fetched. This does not rewrite URLs inside embedded HTML or implement HTTP Content-Location bases.

How are RSS GUIDs handled? RSS GUIDs are opaque identifiers. Attribute-bearing GUIDs use element text, not attribute values. Numeric-looking text stays a string, preserving leading zeroes and large IDs for deduplication. An explicit item link takes precedence. When it is absent, a GUID supplies the article link only if isPermaLink is omitted or exactly true and the GUID is an absolute HTTP/HTTPS URL without credentials. A GUID marked false, an invalid flag, a non-HTTP scheme or a relative reference does not supply a link. GUID destinations are never fetched or DNS-validated. This is not full RSS conformance validation.

Who is this for compared with a feed parsing library?

Content operations and marketing automation teams that need scheduled, batched JSON delivery through an Actor rather than hosting a parser. Universal Feed Parser already supports relative URI resolution and more embedded-content handling; this Actor is a hosted workflow convenience, not a claim of broader parsing coverage. Store keywords: RSS to JSON, Atom normalization, content aggregation, feed monitoring.

How will this offering be evaluated? After a purchasable listing exists, keep if days 61-90 produce at least 30 paid runs, five paying users, positive contribution margin and at least 95% success on supported public feeds. Revisit listing/pricing after 100 qualified impressions with fewer than five paying users. Archive at day 90 if fewer than five paying users and under $10 gross revenue, with no repeat customer or concrete inbound request. These are proposed evaluation thresholds, not measured demand.

Can I use this for authenticated or private feeds? No. This actor only fetches public feeds without credentials.