Markdown to HTML

LLMs default to markdown - headers, bold text, bullet lists, code blocks - because it's what most chat interfaces render. The moment you want to do something with that output outside a chat window - post it to a Discord or Feishu bot (both explicitly mentioned in the pack's README as integration targets), embed it in an email, or render it as an image card - markdown syntax stops being useful and you need real HTML. This node does that one conversion.

How it works

Feed it a markdown string, get back HTML - headers become <h1>/<h2> tags, bold and italic become <strong>/<em>, lists become <ul>/<ol>, and so on, following standard markdown-to-HTML conversion rules.

The inputs and outputs that matter

• md_str - the markdown text to convert. Typically an LLM node's raw output, if the model was asked to format its answer.
• is_enable - the pack's standard bypass toggle.

Output: HTML (STRING) - the converted markup. A natural next step for this output is the pack's own HTML to Image node, if you want to go from "LLM wrote markdown" all the way to "here's a rendered image card" in two steps.

How to install it

Search comfyui_LLM_party in ComfyUI Manager and install, then restart. Manual route:

cd ComfyUI/custom_nodes
git clone https://github.com/heshengtao/comfyui_LLM_party.git

Run pip install -r requirements.txt from the pack's own folder with ComfyUI's Python, then restart. This node needs no API key and no model download - it's a straightforward text transformation, one of the lighter dependencies in a pack whose shared requirements.txt otherwise covers LLM clients, embeddings, OCR, and TTS.

Common issues & troubleshooting

Some markdown doesn't convert the way you expected. Standard markdown (headers, bold/italic, lists, links, code blocks) converts reliably; less common extensions - footnotes, custom containers, some table syntaxes - depend on exactly which markdown library sits behind this node, and support for those isn't guaranteed. If a specific piece of formatting matters, test it directly rather than assuming full CommonMark or GitHub-flavored-markdown coverage.

Output HTML looks unstyled. That's expected - this converts markdown structure into HTML tags, it doesn't add any CSS or visual styling. If you want a styled result, either add your own CSS wrapper around the output before it reaches something like the HTML to Image node, or accept plain, unstyled semantic HTML if that's all the destination (an email client, a bot message) needs.

Nested formatting (bold inside a list item, a link inside a header) doesn't render right. This usually comes down to how the source markdown was written rather than the converter - nested markdown syntax is one of the more fragile corners of the format across every implementation, not specific to this node.