HTML Method

The HTML method extracts feed URLs from HTML content by scanning link elements and anchor tags.

Follows RSS Board Autodiscovery and WHATWG Feed Autodiscovery specs.

How It Works

The HTML method scans two types of elements:

Link Elements

Looks for <link> elements that advertise feeds:

<!-- rel="alternate" with feed MIME type -->
<link rel="alternate" type="application/rss+xml" href="/feed.xml" />

<!-- rel="feed" (WHATWG spec) -->
<link rel="feed" href="/feed" />

Anchor Elements

Scans <a> tags for feed links using two strategies:

1. URI matching: Checks if href ends with a common feed path like /feed or /rss.xml, or has a feed segment like /rss/ in its path.
2. Label matching: Checks if the link text, title, or aria-label contains words like "RSS", "Feed", "Atom". This catches icon-only links that have no visible text.

<!-- Matched by URI -->
<a href="/feed.xml">XML</a>

<!-- Matched by label (text) -->
<a href="/subscribe">RSS Feed</a>

<!-- Matched by label (title / aria-label) on an icon-only link -->
<a href="/blog/syndication" title="RSS feed"><svg>...</svg></a>

Configuration

Feedscout comes with reasonable defaults, but you can customize how HTML is parsed if needed.

Link Selectors

Control which <link> elements are matched:

import { mimeTypes } from 'feedscout/feeds'

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      linkSelectors: [
        { rel: 'alternate', types: mimeTypes },
        { rel: 'feed' },
      ],
    },
  },
})

Anchor URIs

Specify URI patterns to match at the end of anchor href attributes:

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      anchorUris: ['/feed', '/rss', '/atom', '/rss.xml', '/feed.xml'],
    },
  },
})

Anchor Labels

Specify text patterns to match in anchor content:

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      anchorLabels: ['rss', 'feed', 'atom', 'syndication'],
    },
  },
})

Anchor Attributes

Specify element attributes to scan for the anchor labels, on the anchor itself and on its descendants. This finds icon-only feed links whose label lives in an attribute, not in visible text: a title, an aria-label, or the layer name Framer emits on a feed icon (data-framer-name="RSS Icon"):

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      anchorAttributes: ['aria-label', 'title', 'data-framer-name'],
    },
  },
})

Anchor Path Segments

Specify patterns to match against the pathname of anchor href attributes. Only the pathname is checked, so a feed path inside a query string does not count:

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      anchorPathSegments: [/\/rss\//, /\/atom\//, /\/feed\//],
    },
  },
})

Ignored URIs

Exclude certain URI patterns from anchor matching:

const feeds = await discoverFeeds(url, {
  methods: {
    html: {
      anchorIgnoredUris: ['wp-json/oembed/', 'wp-json/wp/'],
    },
  },
})

Default Values

You can import the default HTML options:

import { defaultHtmlOptions } from 'feedscout/feeds'

The defaults include comprehensive anchor URIs and common feed-related labels. You can also import individual pieces:

import {
  linkSelectors,
  anchorLabels,
  anchorAttributes,
  anchorPathSegments,
  urisComprehensive,
  ignoredUris,
} from 'feedscout/feeds'

Using Directly

You can use the HTML discovery function directly to get URIs without validation:

import { discoverUrisFromHtml } from 'feedscout/methods'

const uris = discoverUrisFromHtml(htmlContent, {
  baseUrl: 'https://example.com',
  linkSelectors: [{ rel: 'alternate', types: ['application/rss+xml'] }],
  anchorUris: ['/feed'],
  anchorLabels: ['rss'],
  anchorIgnoredUris: [],
})

// [
//   '/feed.xml',
//   '/rss',
// ]

URIs come back as written in the HTML, so they can be relative. They are resolved against <base href> when the page has one. discoverFeeds resolves the rest against the page URL.