🔀 TA Prompt Hub

Messy workflows hide their prompts all over the graph. TA Prompt Hub is the pack's attempt at a single accounting point: one node where the positive, negative, additional, and LoRA-trigger prompts all pass through, so you can see at a glance what a run is actually being told.

It's deliberately dumb - pure pass-through, no routing. Three required multiline text inputs, positive_prompt, negative_prompt, and additional_prompt (the latter is the "style and scene extras" slot, kept separate so LoRA trigger words don't get mixed into your style text). Plus an optional lora_trigger_words string input for the trigger words a LoRA loader hands you. Each comes back out on its own output, unchanged, and there's a fifth output: combined_prompt, which joins positive + additional + lora triggers with ", ", skipping any that are empty.

That combined output is the quietly useful part. Instead of hand-splicing prompt strings with concat nodes and worrying about trailing commas, you get one string that's ready for a single CLIP Text Encode. It's the "hub" doing what hubs do - collecting everything in one place so the downstream encoder sees one coherent prompt. All four prompt fields support dynamic prompts syntax, so {wildcards} style processing works straight through.

Where this slots into a TA workflow: Prompt Hub gathers the manual/text side, TA Prompt Controller routes or blends it with the LLM side, and the winner feeds the encoder. If you open a shared workflow and want to know what prompts it's running with, the Hub is the first node you'd look at - that's its job, and it does it without any hidden transformation.

Install. No Python dependencies; pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/tmode-1960/TA-ComfyUI-Nodes-Pack

or ComfyUI Manager (search "TA ComfyUI Nodes Pack"), restart, under TA Nodes/text.

Gotchas. It's a pass-through, so there's not much to break - but don't expect it to route anything. That's explicitly TA Prompt Controller's job; the two are a pair, and mixing them up is the most common confusion. lora_trigger_words is optional, so if it's unconnected the combined prompt just omits it. And the usual pack note: TA v1 workflows need rebuilding after the v2.x breaking change before these nodes will be found.