Combine Prompts
Stitch prompt parts together without leaving comma scars
- string
There's a specific annoyance in ComfyUI that Combine Prompts exists to kill: building a prompt out of parts and getting masterpiece, best quality, , 1girl, solo - with a dangling comma where two pieces met. If you've ever piped a tagger's output into a string next to a hand-written prefix, you know the pain. This node is the pack's two-line fix: take a start and an end, join them, and make sure the seam is clean.
It's a pure string utility - no model, no VRAM, nothing but text in and text out - which makes it a handy little building block for the pack's tagger and caption workflows.
How it works
The mechanism is refreshingly honest: the node concatenates the end text after the start text. The useful part is what the author does to start first - the tooltip says it's "normalized for prompt separator." In practice that means the node handles the joining convention for you, so you can pipe in pieces without obsessing over trailing commas. It's the same class of glue logic as the pack's caption-token merging, but aimed at prompts rather than caption files.
start- the text that goes first, normalized so the join doesn't produce, ,seams.end- appended after start.
One output: string, ready to feed a positive/negative conditioning node, a text encoder, or further prompt processing.
When you'd reach for it
Two workflows make this node earn its keep.
The first is tagger pipelines. You run OppaiOracle Tagger or PixAI Tagger, which give you prompt-ready Danbooru-style tag lists, and you want to prepend a quality prefix (masterpiece, best quality) or append a style suffix. Rather than doing string surgery in a Text Concatenate and hunting comma bugs, you wire the prefix into start and the tag list into end and move on.
The second is prompt-from-parts architecture: a shared base prompt, plus a per-image variation fed in as end. Because start is normalized for the separator, swapping the variation doesn't leave broken punctuation behind. Combined with the pack's Remove Prompt Comments and its wildcard/template node, you can assemble prompts from clean, composable pieces and never hand-edit a full string again.
The honest caveat
This is a small node, and if you're already comfortable with ComfyUI's built-in text concatenation you might not miss it - plain concatenation works fine when you control both pieces. Where it wins is exactly the seam handling and the consistency it gives a graph you share or re-open months later. It's also not a formatter: it won't sort tags, deduplicate them, or clean up inside either input. That's what the pack's caption-token utilities are for.
Installing
Part of kinorax/comfyui-info-prompt-toolkit. Install via ComfyUI Manager (search "Info-Prompt-Toolkit") or:
cd ComfyUI/custom_nodes
git clone https://github.com/kinorax/comfyui-info-prompt-toolkit.git
cd comfyui-info-prompt-toolkit
pip install -r requirements.txt
Restart ComfyUI. No model files, no extra dependencies - it's pure string handling, so it's one of the lowest-risk nodes in the pack to add to an existing install. If you're already running the pack for its taggers or metadata tools, this is a freebie you'll find yourself dropping into prompt graphs more often than you'd expect.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| startopt | STRING | Text placed at the beginning (normalized for prompt separator) | |
| endopt | STRING | Text appended after start |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| string | STRING | — |