RSS Utils

RSS Utils adds a focused set of Flow-Like nodes for working with RSS, Atom, RDF, and JSON Feed content. Use it to discover feed links on web pages, prepare feed HTTP requests, parse feed documents into a consistent structure, filter and deduplicate items, and turn feed entries into Markdown for digests or notifications.

What You Can Build

• Monitor product updates, blogs, changelogs, news sites, GitHub releases, and community feeds.

• Create daily or weekly content digests from one or more feeds.

• Filter feed items by release date, category, or text search.

• Normalize feeds from different publishers into one shared item format.

• Remove duplicate items before sending notifications or storing results.

• Convert feed items into Markdown for summaries, emails, chat messages, notes, or downstream AI workflows.

Included Nodes

Normalize Feed URL

Cleans up user-supplied feed URLs before they are used in a flow.

Use this when feed URLs may come from forms, documents, copied browser addresses, or generated text. The node trims the input, can add https:// when the scheme is missing, and removes fragments that are not useful for fetching feeds.

Build Feed Request

Creates a structured HTTP request descriptor for fetching feeds.

The node sets feed-friendly request headers, including an Accept header for RSS, Atom, JSON Feed, XML, and JSON responses. It also supports custom User-Agent, ETag, Last-Modified, and timeout settings, so scheduled flows can perform conditional fetches and avoid unnecessary downloads when a feed has not changed.

This node prepares the request. Pair it with an HTTP fetch node in your flow to perform the actual network request.

Parse Feed

Parses feed document text into normalized feed metadata and normalized feed items.

Supported input formats include RSS, Atom, RDF, and JSON Feed. The parser returns the detected feed type, item count, warnings, the full parsed feed object, and the item array that most downstream nodes consume.

This is the central parser for the package. Different publishers use feed fields in different ways, so this node normalizes common item data such as title, link, id, author, categories, summary, content, publication date, updated date, and release date into one predictable structure.

Extract Feed Links

Finds feed links inside HTML pages.

Use this when you have a normal website URL and need to discover available RSS, Atom, or JSON Feed URLs. The node scans HTML for alternate feed links and resolves relative links against the page URL.

Search Feed Items

Keeps only feed items whose text contains a query.

The search is case-insensitive and checks common item fields such as title, link, summary, content, and author. Use it to keep only matching posts, releases, alerts, or articles after parsing a feed.

Filter Feed Items by Category

Keeps only feed items with a matching category.

The category comparison is case-insensitive. Use this for feeds that tag items by product, topic, team, section, repository, region, or content type.

Filter Feed Items by Date

Keeps only feed items inside an inclusive release-date window.

This node is designed for scheduled workflows and digest generation. For example, you can keep only items released today, this week, after the last run, or before a cutoff date.

The node uses Flow-Like date pins and also accepts feed-style date text. It can handle common feed date formats including RFC3339, RFC2822, HTTP-date, Unix timestamps, YYYY, YYYY-MM, YYYY-MM-DD, and named dates such as June 4, 2026. Date-only values expand to the whole day, month, or year, so a filter like 2026-06-04 includes all matching items on that calendar day.

You can choose which item date to compare:

• release_date: the package's normalized best release date.

• published: the item publication date.

• updated: the item update date.

• any: match against any available item date.

An optional setting controls whether undated items should be kept when a date filter is active.

Dedupe Feed Items

Removes repeated feed items.

Feeds often repeat content through mirrors, updated entries, alternate URLs, or combined subscriptions. This node removes duplicates using stable identifiers when available, then falls back to link and title/date based matching.

Use this before limiting, formatting, notifying, or storing feed items.

Limit Feed Items

Keeps the first N feed items.

Use this after parsing, filtering, or deduplication to cap the amount of content that continues through the flow. This is useful for dashboards, digests, notifications, and AI summarization steps where you want predictable output size.

Feed Item to Markdown

Formats one normalized feed item as Markdown.

The node can include the item link, summary, and full content depending on your flow. Use it to prepare readable output for digests, chat messages, notes, reports, or prompt inputs.

Example Workflows

Website Feed Discovery

1. Fetch a website homepage.

1. Use Extract Feed Links to find available feeds.

1. Normalize the selected URL with Normalize Feed URL.

1. Build a fetch request with Build Feed Request.

1. Fetch the feed body and parse it with Parse Feed.

Daily RSS Digest

1. Fetch and parse one or more feeds.

1. Use Filter Feed Items by Date to keep items from the current day.

1. Use Dedupe Feed Items to remove repeated posts.

1. Use Limit Feed Items to cap the digest size.

1. Convert each item with Feed Item to Markdown.

1. Send the Markdown to an email, chat, notes, or summary workflow.

Product Update Monitor

1. Parse vendor changelog, cloud provider, GitHub, or product feeds.

1. Use Search Feed Items for keywords such as product names, service names, vulnerability terms, or release channels.

1. Use Filter Feed Items by Category when the feed provides useful tags.

1. Keep only recent items with Filter Feed Items by Date.

1. Send matched items to a notification or triage workflow.

Feed Compatibility

The package is tested against a broad corpus of public feed sources, including Reddit feeds, AWS feeds, news feeds, technology feeds, and GitHub Atom feeds. The normal test suite runs offline and validates parser behavior and node coverage without depending on live third-party services.

Notes

• This package focuses on feed utilities and parsing. It does not include a network fetch node; use your existing HTTP node with Build Feed Request.

• Date filtering is inclusive, which is usually what scheduled feed workflows need.

• Parser warnings are exposed so flows can inspect malformed or unusual feeds without losing successfully parsed items.

License

Licensed under either the MIT License or the Apache License, Version 2.0, at your option.

Copyright (c) 2026 Rheosoph GmbH, Felix Schultz.