RSS Feed Parser (WebAPI)

RSS is the quiet backbone of a lot of automation, and ComfyUI can talk to it like anything else on the network. RSS Feed Parser (WebAPI) fetches a feed URL and returns its entries as a list - title, link, description, publish date - so you can pipe "what's new on this blog/forum/release page" into a prompt, a workflow, or a save node. It's an odd tool to see in an image-generation pack, and it's exactly the kind of odd tool that makes automation fun.

The default feed is Python Insider's, which tells you where this node came from - someone's release-watching workflow. Point it at a release feed for the model you use and you have a "did it update?" primitive.

How it works

It fetches the feed with requests and parses it with regex, not a full XML parser: it finds <item> blocks and pulls the title, link, description, and pubDate out of each. max_items (1–50, default 5) caps how many entries come back, and the optional filter_keyword restricts results to entries whose title or summary contains that word, case-insensitively.

Outputs are entries (a LIST of entry dicts) and raw_xml (the fetched feed text, handy when regex missed something and you want to see the source).

The gotcha, and it's a real one

Regex-based parsing means this is an RSS-format parser. It looks for <item> blocks - that's the RSS 2.0 shape. Atom feeds, which use <entry> instead, will parse to an empty list, despite the node's description mentioning Atom. For a well-formed RSS 2.0 feed it's fine; for Atom, heavily-namespaced feeds, or malformed XML, expect zero results and reach for the raw_xml output to see why. If RSS compliance matters to you, a real XML parser is more robust - this is the quick-and-dirty version.

Also, it's a Web API node, so it needs requests, which ComfyUI doesn't bundle.

Installing it

cd ComfyUI/custom_nodes/
git clone https://github.com/TensorVizion/OmniNodes
cd OmniNodes
pip install -r requirements.txt

Then restart ComfyUI (ComfyUI Manager: search "OmniNodes"), installing into the same Python environment ComfyUI runs in. Skip the pip line and you'll get ModuleNotFoundError: No module named 'requests' on import. Registers under TensorVizion/Web API.

The natural full loop in this pack: RSS Feed Parser → JSON Field Extractor to pull a headline → Prompt Combiner to build a prompt from it → sampler → Response Saver to log the result. It's a strange but genuinely fun way to make your ComfyUI machine read the internet.