Doom Text Show

The most underrated debugging tool in any node pack is the one that just shows you the text. Doom Text Show takes a string on its text input, displays it in the node's UI so you can actually read it, and passes the same string out through text_output unchanged. That's it - and that's genuinely valuable the moment your prompt comes from anywhere other than your own typing.

Think about the Doom pack's own pipelines: DoomRandomPrompt assembles a prompt from wildcards and category files. DoomMultiPromptLoader pulls one from a file and advances an index. DoomPromptBuilder stitches seven sections together. Doom Translate rewrites your words into another language. Any one of those can produce something you did not expect, and the fastest way to find out is a node that shows you the result before you burn a full generation on it.

How it works

Drop a Doom Text Show at the end of any string chain - prompt sources, save-node output_message receipts, translate output, whatever - and the text renders in the node body. Because text_output is a faithful pass-through, you can insert it mid-chain without changing what flows downstream. It's an output node by nature (the brief marks it is_output_node: true), but because it also passes the value through, it plays nicely inline too.

The inputs and outputs

• text - the string to display. Usually wired, not typed.
• text_output - the same string, for continuing the chain.

That's the whole surface. If you want a node that's only a display and ends the branch, this qualifies; if you want to see the value and keep it flowing, it does that too.

Setup

Standard pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/PeterMikhai/Doom_Flux_NodePack

Restart ComfyUI, or search "DoomAI Nodes" in Manager. No dependencies.

Where people get burned

• Expecting an editable box. Doom Text Show is a display. To enter text, use Doom Text Enter (they're deliberately a pair: enter vs. show).
• Forgetting it's mid-chain-safe. Since the output is a passthrough, some people skip adding it because they think they'll have to re-wire. You won't - insert it and the chain keeps working.
• The UI text is what the node renders. If a prompt source gives you something with wildcards still unexpanded (e.g. you see __file__ in the output), the show node just displays what it got - the problem is upstream, and this node is precisely how you find that out.

In a pack full of loaders that need 40 GB of weights and samplers that need 24 GB of VRAM, the most useful node on a bad day might be the free green one that shows you your own text. Cheap, obvious, and you'll reach for it constantly.