Taglist Prompt

Here's the honest truth about this node: right now, it is the Safe Tag List Prompt. The pack's README says it plainly - "Direct tag list prompt node that inherits the SafeTagListPromptNode behavior" - and the source is literally class TagListPromptNode(SafeTagListPromptNode) with a different category label and display name. Same inputs, same outputs, same blocklist, same safe_mode defaulting to true. So if you already read that article, you know everything this node does.

So why does it exist? The reasonable guess is that it's the pack author's intended unsafe sibling - the "Taglist" (without "Safe") that was supposed to generate explicit taglists - but it currently just inherits the safe behavior wholesale. Either the split is half-built or the author decided the safe defaults were good enough for both. Either way, as of this writing, they're interchangeable.

How it works (identical to Safe Tag List Prompt)

You paste a working template_text taglist and a custom_idea. With safe_mode on, the node checks both against the blocklist (the full default NSFW vocabulary), rejects the input if either trips it, then sends the idea to Grok with a system prompt asking for a taglist that mirrors the template's structure and density, kept SFW. The response gets a second blocklist check before it's returned as taglist_prompt. Temperature defaults to 0.2, model to grok-3-latest.

The inputs that matter

Same as the Safe node:

• template_text - required, the reference taglist whose shape gets copied.
• custom_idea - required, the new concept.
• api_key - required; no env-var fallback.
• safe_mode - on by default (worth checking if you expected this node to be the "raw" one - it isn't).
• blocklist - editable vocabulary, defaulting to the SFW blocklist.

Output: one STRING, taglist_prompt.

Install

ComfyUI Manager → babydjacNODES, or git clone https://github.com/babydjac/babydjacNODES into ComfyUI/custom_nodes, restart, hard-refresh. No extra pip deps.

Gotchas

• Do not assume this is the uncensored variant. The name suggests it, the README's phrasing hints at it, but the code says otherwise. If you flip safe_mode off, you still get the same Grok call the Safe node makes - just without the blocklist guards.
• Two nearly identical nodes in your node menu. Search "taglist" and you'll see both "Safe Tag List Prompt" and "Taglist Prompt." Pick one and be consistent; there's no functional difference today.
• Same cost model - one paid Grok call per run, same as its sibling.

Honestly, the most useful thing you can do with this node is decide whether you even need it. If the Safe variant does the job, use that one and bypass this; if the author later splits this into a genuinely explicit prompt generator, then it becomes interesting. For now it's a duplicate with a slightly misleading name.