APNext Brands

This one's exactly what it sounds like: a dropdown of 290 brand names - Adidas, Adobe, Airbnb, all the way down - that it staples onto your prompt. It's one of the more niche members of dagthomas's "SDXL Auto Prompter" pack, and I'll be straight with you, it's more of a novelty than a daily driver. The honest use is mood-and-aesthetic: "in the style of a Nike ad," a Balenciaga-campaign look, the glossy product-photography vibe that a brand name evokes. What it is not is a reliable logo generator - diffusion models mangle text and trademarks, and naming a brand nudges palette and styling far more than it reproduces anything official.

Where it can actually earn its keep is advertising mockups and product-viz exploration, where you want the feel of a brand's visual language without hand-writing it. Batch a few random picks and you'll stumble onto aesthetic combinations you wouldn't have thought to type.

How it works

Pure Python, no model, no API key. It joins your prompt text with the brand you pick using the separator (comma by default) and returns a string. Like the rest of this node family, each menu opens with None / Random / Multiple Random, so with a seed set you can let it roll brands for you. It's a tag-style injection, so it behaves best on SDXL-lineage checkpoints; on 2026 LLM-encoded models the name still goes in as text but reads as instruction rather than a weighted tag.

The inputs and outputs that matter

• prompt - your base text; the brand gets appended.
• brands - the 290-entry dropdown. This is the whole point of the node.
• seed - drives the Random / Multiple Random rolls for batch variety.
• attributes (off by default) - folds in the extra descriptor words the pack attaches to entries.
• string - an optional prefix string you can prepend for styling.

Two string outputs, prompt and random, into a CLIP Text Encode.

How to install it

ComfyUI Manager: search comfyui_dagthomas, install, restart. Manual route:

cd ComfyUI/custom_nodes && git clone https://github.com/dagthomas/comfyui_dagthomas
cd comfyui_dagthomas && pip install -r requirements.txt

then restart. The requirements.txt installs the whole pack's stack - openai, anthropic, google-generativeai, transformers, decord, scipy - because dagthomas also ships cloud-LLM and video nodes in the same repo. This node needs none of it, but you get it anyway; decord is the dependency most likely to fail a build, and if it does, the plain builder nodes usually still load.

Where people get burned

• Expecting logos. It won't give you a clean, correct trademark. It gives you a vibe. If you need the actual mark, composite it in afterward.
• One brand, not five. Stack a handful of brands and their aesthetics cancel out. Pick one, or use Multiple Random purely as an idea generator and keep the winners.
• Model era. On SDXL/Illustrious this reads as the tag it is. On Flux 2 / Z-Image / Qwen, a single brand name as part of a clean sentence works; a brand pile-up doesn't.
• Diminishing returns. Of all the builders in this pack, this is the one with the narrowest real use. If you're not doing ad-style or product work, the Cinematic or Art builders will move your image more.