ts-rss-builder

A Zero-Dependency TypeScript RSS builder for Astro.

Purpose

Generate RSS feeds from Astro collections with ease. Built for NN1 Spotlights RSS feed.

Why?

For learning and to keep things simple. Most RSS builders have alot of deps and stuff you don't need. My library is zero-dep, easy to understand, and made to play nicely with Astro collections.

Installation

pnpm install ts-rss-builder

Usage

import { getCollection } from 'astro:content';
import { createRSSFromAstroCollection } from 'ts-rss-builder';
import sanitizeHtml from 'sanitize-html';

const baseUrl = "https://example.com";
const blog = await getCollection('blog');

return createRSSFromAstroCollection(blog, {
      site: baseUrl,
      feedUrl: `${baseUrl}/blog.xml`,
      mappers: {
        title: (entry) => entry.data.name,
        link: (entry) => `${baseUrl}/blog/${entry.id}`,
        description: (entry) => generateDescription(entry),	// You'd usually get this from entry.data
        content: (entry) => sanitizeContent(entry.rendered?.html || entry.body || ''), // sanitizeHtml is a custom defined wrapping sanitizeHtml
        pubDate: (entry) => entry.data.date,
        guid: (entry) => `${baseUrl}/blog/${entry.id}`,
        enclosure: (entry) => createImageEnclosure(entry, baseUrl, isDev), // // You'd usually get this from entry.data (image)
      },
});

Configuration Interface

interface AstroRSSConfig<T = any> {
  title: string;                // Feed title
  description: string;          // Feed description
  site: string;                 // Base site URL (e.g. https://example.com)
  feedUrl: string;              // Full URL to the RSS feed (e.g. https://example.com/feed.xml)
  language?: string;            // Optional language code (e.g. "en-GB")
  customData?: string;          // Optional raw XML string for additional RSS tags/namespaces

  mappers: {
    title: (entry: T) => string;              // Map entry to item title
    link: (entry: T) => string;               // Map entry to item link
    description: (entry: T) => string;        // Map entry to item description (typically summary)
    content: (entry: T) => string;             // Map entry to item full content (for content:encoded)
    pubDate: (entry: T) => string | Date;     // Map entry to publication date (ISO string or Date)
    category?: (entry: T) => string | string[] | undefined; // Optional categories/tags
    guid?: (entry: T) => string;               // Optional unique identifier for the item
    enclosure?: (entry: T) => RSSEnclosure | undefined;    // Optional media enclosure
  };
}

interface RSSEnclosure {
  url: string;
  length?: number | string;
  type?: string; // MIME type, e.g. "image/jpeg"
}

Security, XML Safety & Compliance

This builder is designed with XML correctness and security in mind. Here are some techniques it uses:

XML Character Escaping:

• All user-supplied text (titles, descriptions, links, etc.) is escaped to replace characters like &, <, >, ", and ' with their XML entity equivalents (&, <, etc.). -This prevents XML injection, broken markup, and parsing errors.

Safe CDATA Handling:

• Content inside <content:encoded> is wrapped in CDATA sections to allow HTML without heavy escaping.
• This library escapes any ]]> sequences inside content to ]]>, preventing early CDATA section termination, which would corrupt the XML.
• It also removes dangerous control characters (non-printable ASCII) which can cause XML parsers to fail or behave unexpectedly.

Date Formatting:

• Publication dates are strictly formatted to RFC-822 compliant strings, ensuring compatibility with RSS readers and feed validators.
• Enclosure & Category Safety

Caching:

• XML escaping results are cached for performance when escaping repeated strings.