macieklamberski/feedsmith: Robust and fast parser and generator for RSS, Atom, JSON Feed, and RDF feeds, with support for Podcast, iTunes, Dublin Core, and OPML files.

Robust and fast JavaScript parser and generator for RSS, Atom, JSON Feed, and RDF feeds, with support for popular namespaces and OPML files. It provides both universal and format-specific parsers that maintain the original feed structure while offering helpful normalization.

Feedsmith maintains the original feed structure in a clean, object-oriented format. It intelligently normalizes legacy elements, providing you with complete access to all feed data without compromising simplicity.

Features
  ·  
Installation
  ·  
Parsing
  ·  
Generating
  ·  
Benchmarks
  ·  
FAQ

• Normalizes legacy elements ✨ — Upgrades feed elements to their modern equivalents so that you never need to worry about reading feeds in older formats.
• CaSe INSENsiTive — Handles fields and attributes in any case (lowercase, uppercase, mixed).

• Fast parsing — One of the fastest feed parsers in JavaScript (see benchmarks).
• Type-safe API — TypeScript type definitions are available for each feed format, making it easy to work with the data.
• Tree-shakable — Only include the parts of the library you need, reducing bundle size.
• Well-tested — Comprehensive test suite with 1200+ tests and 99% code coverage.

• Works in Node.js and all modern browsers.
• Works with plain JavaScript, you don’t need to use TypeScript.

✅ Available
  ·  
⌛️ Work in progress
  ·  
📋 Planned

The easiest way to parse any feed is to use the universal parseFeed function:

import { parseFeed } from 'feedsmith'

const { type, feed } = parseFeed('feed content')

console.log('Feed type:', type) // → rss, atom, json, rdf
console.log('Feed title:', feed.title)

if (type === 'rss') {
  console.log('RSS feed link:', feed.link)
}

If you know the format in advance, you can use the format-specific parsers:

import { parseAtomFeed, parseJsonFeed, parseRssFeed, parseRdfFeed } from 'feedsmith'

// Parse the feed content
const atomFeed = parseAtomFeed('atom content')
const jsonFeed = parseJsonFeed('json content')
const rssFeed = parseRssFeed('rss content')
const rdfFeed = parseRdfFeed('rdf content')

// Then read the TypeScript suggestions for the specific feed type
rssFeed.title
rssFeed.dc?.creator
rssFeed.dc?.date
rssFeed.sy?.updateBase
rssFeed.items?.[0]?.title

Parsing OPML files is as simple:

import { parseOpml } from 'feedsmith'

// Parse the OPML content
const opml = parseOpml('opml content')

// Then read the TypeScript suggestions
opml.head?.title
opml.body?.outlines?.[0].text
opml.body?.outlines?.[1].xmlUrl

The objects returned from the parser functions are highly comprehensive, aiming to recreate the actual feed structure and its values, including all the supported namespaces. Below are some examples of what is available.

import { parseAtomFeed } from 'feedsmith'

const atomFeed = parseAtomFeed(`
  
  
    
    example-feed
    John Doe
    Jane Smith
    2022-01-01T12:00+00:00
    This is an example of description.
    2000-01-01T12:00+00:00
    hourly
    1
    
      
      example-entry
      Jack Jackson
      2022-01-01T12:00+00:00
    
  
`)

atomFeed.title // → Example Feed
atomFeed.dc?.contributor // → Jane Smith
atomFeed.dc?.date // → 2022-01-01T12:00+00:00
atomFeed.sy?.updateFrequency // → 1
atomFeed.entries?.[0]?.title // → Example Entry
atomFeed.entries?.[0]?.dc?.creator // → Jack Jackson

Returns:

{
  "id": "example-feed",
  "title": "Example Feed",
  "entries": [
    {
      "id": "example-entry",
      "title": "Example Entry",
      "dc": {
        "creator": "Jack Jackson",
        "date": "2022-01-01T12:00+00:00"
      }
    }
  ],
  "dc": {
    "creator": "John Doe",
    "description": "This is an example of description.",
    "contributor": "Jane Smith",
    "date": "2022-01-01T12:00+00:00"
  },
  "sy": {
    "updatePeriod": "hourly",
    "updateFrequency": 1,
    "updateBase": "2000-01-01T12:00+00:00"
  }
}

import { parseRssFeed } from 'feedsmith'

const rssFeed = parseRssFeed(`
  
  
    
      
        http://example.org/banner.png
        http://example.org/
        Quos placeat quod ea temporibus ratione
        80
        15
      
      
        
        
        q
        http://example.org/mt/mt-search.cgi
      
      
        0
        20
        21
        22
        23
      
      
        Monday
        Wednesday
        Friday
      
      
        
        http://example.org/item/1
        Some description of the first item.
        http://example.org/comments/1
        
        http://example.org/guid/1
        Thu, 05 Sep 2002 0:00:01 GMT
        Example's Realm
      
    
  
`)

rssFeed.title // → Sample Feed
rssFeed.textInput?.description // → Search this site:
rssFeed.items?.length // → 1
rssFeed.items?.[0]?.enclosure?.url // → http://example.org/audio/demo.mp3

{
  "title": "Sample Feed",
  "link": "http://example.org/",
  "description": "For documentation only",
  "language": "en",
  "webMaster": "webmaster@example.org",
  "pubDate": "Sat, 19 Mar 1988 07:15:00 GMT",
  "lastBuildDate": "Sat, 19 Mar 1988 07:15:00 GMT",
  "categories": [{ "name": "Examples2", "domain": "http://www.example.com/cusips" }],
  "generator": "Sample Toolkit",
  "docs": "http://feedvalidator.org/docs/rss2.html",
  "cloud": {
    "domain": "rpc.example.com",
    "port": 80,
    "path": "/RPC2",
    "registerProcedure": "pingMe",
    "protocol": "soap"
  },
  "ttl": 60,
  "image": {
    "url": "http://example.org/banner.png",
    "title": "Example banner",
    "link": "http://example.org/",
    "description": "Quos placeat quod ea temporibus ratione",
    "height": 15,
    "width": 80
  },
  "textInput": {
    "title": "Search",
    "description": "Search this site:",
    "name": "q",
    "link": "http://example.org/mt/mt-search.cgi"
  },
  "skipHours": [0, 20, 21, 22, 23],
  "skipDays": ["Monday", "Wednesday", "Friday"],
  "items": [
    {
      "title": "First item title",
      "link": "http://example.org/item/1",
      "description": "Some description of the first item.",
      "comments": "http://example.org/comments/1",
      "enclosure": {
        "url": "http://example.org/audio/demo.mp3",
        "length": 1069871,
        "type": "audio/mpeg"
      },
      "guid": "http://example.org/guid/1",
      "pubDate": "Thu, 05 Sep 2002 0:00:01 GMT",
      "source": { "title": "Example's Realm", "url": "http://www.example.org/links.xml" }
    }
  ]
}

import { parseOpml } from 'feedsmith'

const opml = parseOpml(`
  
  
    
      
      Mon, 15 Jan 2024 09:45:30 GMT
      Jack Smith
    
    
      
      
    
  
`)

opml.head?.title // → Tech Sites
opml.body?.outlines?.[0].text // → The Verge
opml.body?.outlines?.[1].xmlUrl // → https://techcrunch.com/feed/

For more examples, check the */references folders in the source code. There, you’ll find the complete objects returned from the parser functions for the various feed formats and versions.

If the feed is unrecognized or invalid, an Error will be thrown with a descriptive message.

import { parseFeed, parseJsonFeed } from 'feedsmith'

try {
  const universalFeed = parseFeed('')
} catch (error) {
  // Error: Unrecognized feed format
}

try {
  const jsonFeed = parseJsonFeed('{}')
} catch (error) {
  // Error: Invalid feed format
}

You can detect feed formats without parsing them.

import { detectAtomFeed, detectJsonFeed, detectRssFeed, detectRdfFeed } from 'feedsmith'

if (detectAtomFeed(content)) {
  console.log('This is an Atom feed')
}

if (detectJsonFeed(content)) {
  console.log('This is a JSON feed')
}

if (detectRssFeed(content)) {
  console.log('This is an RSS feed')
}

if (detectRdfFeed(content)) {
  console.log('This is an RDF feed')
}

Although JSON feeds are simply JSON objects that can be easily generated manually, the generateJsonFeed function provides helpful type hints, which can aid in feed generation. Additionally, you can use Date objects for dates, which are automatically converted to the correct format in the background.

import { generateJsonFeed } from 'feedsmith'

const jsonFeed = generateJsonFeed({
  title: 'My Example Feed',
  feed_url: 'https://example.com/feed.json',
  authors: [
    {
      name: 'John Doe',
      url: 'https://example.com/johndoe',
    },
  ],
  items: [
    {
      id: '1',
      content_html: '
Hello world
',
      url: 'https://example.com/post/1',
      title: 'First post',
      date_published: new Date('2019-03-07T00:00:00+01:00'),
      language: 'en-US',
    },
  ],
})

Will output:

{
  "version": "https://jsonfeed.org/version/1.1",
  "title": "My Example Feed",
  "feed_url": "https://example.com/feed.json",
  "authors": [
    {
      "name": "John Doe",
      "url": "https://example.com/johndoe",
    },
  ],
  "items": [
    {
      "id": "1",
      "content_html": "
Hello world
",
      "url": "https://example.com/post/1",
      "title": "First post",
      "date_published": "2019-03-06T23:00:00.000Z",
      "language": "en-US",
    },
  ],
}

import { generateOpml } from 'feedsmith'

const opml = generateOpml({
  head: {
    title: 'My Feed',
    dateCreated: new Date(),
  },
  body: {
    outlines: [
      {
        text: 'My Feed',
        type: 'rss',
        xmlUrl: 'https://example.com/feed.xml',
        htmlUrl: 'https://example.com',
      },
    ],
  },
})

Will output:

xml version="1.0" encoding="utf-8"?>
<opml version="2.0">
  <head>
    <title>My Feedtitle>
    <dateCreated>Fri, 11 Apr 2025 13:05:26 GMTdateCreated>
  head>
  <body>
    <outline text="My Feed" type="rss" xmlUrl="https://example.com/feed.xml" htmlUrl="https://example.com"/>
  body>
opml>

A comprehensive set of benchmarks, categorized by various file sizes, is available in the /benchmarks directory. These benchmarks were conducted using both Tinybench and Benchmark.js.

See full benchmark results →

For a quick overview, here are the results of parsing RSS, Atom, and RDF feeds using various JS packages with Tinybench. Feedsmith’s results are marked with an asterisk (*).

📊 RSS feed parsing (50 files × 100KB–5MB)
┌───┬───────────────────────────────┬─────────┬──────────────┬──────────┬──────────┬──────┐
│   │ Package                       │ Ops/sec │ Average (ms) │ Min (ms) │ Max (ms) │ Runs │
├───┼───────────────────────────────┼─────────┼──────────────┼──────────┼──────────┼──────┤
│ 0 │ feedsmith *                   │ 7.34    │ 136.167      │ 128.479  │ 173.223  │ 111  │
│ 1 │ @rowanmanning/feed-parser     │ 7.16    │ 139.678      │ 128.722  │ 170.903  │ 108  │
│ 2 │ @ulisesgascon/rss-feed-parser │ 4.14    │ 241.405      │ 230.806  │ 278.534  │ 63   │
│ 3 │ feedparser                    │ 2.50    │ 399.824      │ 374.049  │ 459.730  │ 38   │
│ 4 │ @extractus/feed-extractor     │ 2.26    │ 443.065      │ 430.349  │ 460.195  │ 34   │
│ 5 │ feedme.js                     │ 2.05    │ 487.222      │ 443.837  │ 535.029  │ 31   │
│ 6 │ rss-parser                    │ 1.66    │ 603.044      │ 573.516  │ 653.683  │ 25   │
│ 7 │ @gaphub/feed                  │ 0.94    │ 1068.621     │ 995.044  │ 1138.913 │ 15   │
└───┴───────────────────────────────┴─────────┴──────────────┴──────────┴──────────┴──────┘

📊 Atom feed parsing (50 files × 100KB–5MB)
┌───┬───────────────────────────┬─────────┬──────────────┬──────────┬──────────┬──────┐
│   │ Package                   │ Ops/sec │ Average (ms) │ Min (ms) │ Max (ms) │ Runs │
├───┼───────────────────────────┼─────────┼──────────────┼──────────┼──────────┼──────┤
│ 0 │ feedsmith *               │ 0.98    │ 1020.035     │ 998.660  │ 1084.180 │ 15   │
│ 1 │ @gaphub/feed              │ 0.95    │ 1058.126     │ 989.001  │ 1150.486 │ 15   │
│ 2 │ @rowanmanning/feed-parser │ 0.63    │ 1580.462     │ 1563.357 │ 1607.379 │ 10   │
│ 3 │ feedparser                │ 0.37    │ 2687.488     │ 2624.427 │ 2751.504 │ 6    │
│ 4 │ @extractus/feed-extractor │ 0.32    │ 3136.880     │ 3107.170 │ 3228.099 │ 5    │
│ 5 │ feedme.js                 │ 0.26    │ 3812.545     │ 3759.928 │ 3843.974 │ 4    │
│ 6 │ rss-parser                │ 0.18    │ 5539.014     │ 5479.560 │ 5609.397 │ 3    │
└───┴───────────────────────────┴─────────┴──────────────┴──────────┴──────────┴──────┘

📊 RDF feed parsing (50 files × 100KB–5MB)
┌───┬───────────────────────────┬─────────┬──────────────┬──────────┬──────────┬──────┐
│   │ Package                   │ Ops/sec │ Average (ms) │ Min (ms) │ Max (ms) │ Runs │
├───┼───────────────────────────┼─────────┼──────────────┼──────────┼──────────┼──────┤
│ 0 │ @rowanmanning/feed-parser │ 13.52   │ 73.990       │ 69.404   │ 89.504   │ 203  │
│ 1 │ feedsmith *               │ 10.16   │ 98.396       │ 92.418   │ 118.053  │ 153  │
│ 2 │ @extractus/feed-extractor │ 3.83    │ 260.946      │ 252.991  │ 274.432  │ 58   │
│ 3 │ feedparser                │ 1.96    │ 509.686      │ 494.823  │ 530.224  │ 30   │
│ 4 │ feedme.js                 │ 1.40    │ 714.442      │ 661.440  │ 789.395  │ 22   │
│ 5 │ rss-parser                │ 0.97    │ 1028.245     │ 985.521  │ 1107.122 │ 15   │
│ 6 │ @gaphub/feed              │ 0.97    │ 1031.579     │ 1008.220 │ 1060.322 │ 15   │
└───┴───────────────────────────┴─────────┴──────────────┴──────────┴──────────┴──────┘

The key advantage of Feedsmith is that it preserves the original feed structure exactly as provided in each specific feed format.

Many alternative packages attempt to normalize data by:

• Merging distinct fields like author, dc:creator, and creator into a single property.
• Combining date fields such as dc:date and pubDate without preserving their sources.
• Handling multiple elements inconsistently, sometimes keeping only the first or last one or ignoring different rel attributes.
• Some libraries try to combine different feed formats into one universal structure.

While this approach can be useful for quick reading of feed data, it often results in a loss of information that may be crucial for certain applications, such as reading data from specific namespaces.

In the course of parsing hundreds of thousands of feeds, I have found that dates in feeds use many different formats. Rather than attempting to parse them all (and potentially introducing errors), dates are returned in their original string form. This approach allows you to use your preferred date parsing library or simply the Date object.

Feedsmith focuses on parsing feeds rather than validating them. It will extract whatever valid data it can find, even from partially valid feeds. This approach makes it more resilient when dealing with feeds found in the wild.

It will only fail if the feed is completely invalid or it does not contain all the fields required according to the specification.

Feedsmith is designed to be forgiving. It will extract whatever valid data it can find and ignore missing or invalid elements. This makes it suitable for use with real-world feeds that may not strictly follow specifications.

Even though Feedsmith is more suited for the Node.js environments, it was also tested in modern browsers where it works seamlessly. It’s provided as an ES module.

Licensed under the MIT license.
Copyright 2025 Maciej Lamberski