Higgsfield Model Info

Of the four nodes in the ComfyUI-Higgsfield-Direct pack, this is the boring one - and that's a compliment. Higgsfield Model Info doesn't generate anything, doesn't call the API, and doesn't need a key. It's a local cheat-sheet that dumps the list of models the pack knows about, so you can see what you're working with before you start spending money on generations.

It exists because the pack's other nodes hide their model lists inside dropdowns. If you want to know the exact model ID for seedream/v4/edit - say, to reference it in a workflow, a note, or documentation - this node prints them all out. It's the kind of utility you wire up once, look at, and never think about again, which is honestly all it should be.

What it does

There's exactly one input: category, a dropdown with text-to-image, image-edit, image-to-video, and all (the default). Pick one and the node spits out a single model_info string that lists the models in that category, each with its human name and its full API ID.

It's a pure lookup against the model registry baked into the pack's code - no network, no latency, no cost. You can display the output with a text node, save it, or just read it off the node.

The honest caveat: this list is whatever the pack shipped with, not a live query of Higgsfield's servers. If Higgsfield adds or retires a model, the pack needs an update before Model Info will reflect it. So treat it as "what this version can reach," and if you're chasing a brand-new model, check the pack's GitHub for a newer release first.

Installation

Same shared install as the rest of the pack - there's nothing special about this node:

cd ComfyUI/custom_nodes
git clone https://github.com/jeremieLouvaert/ComfyUI-Higgsfield-Direct.git
pip install -r ComfyUI-Higgsfield-Direct/requirements.txt

or use ComfyUI Manager and search "Higgsfield". Restart, and it's under AKURATE/Higgsfield. Since the node never talks to the API, you can play with it before you've even set up an API key - a nice way to confirm the pack loaded correctly before you commit to any paid generation.

The actual list

Roughly what you'll see, per category:

• Text-to-image: Soul 2.0 (higgsfield-ai/soul/standard), Reve (reve/text-to-image), Seedream 4 (bytedance/seedream/v4/text-to-image)
• Image edit: Seedream 4 Edit (bytedance/seedream/v4/edit)
• Image-to-video: DOP Preview (higgsfield-ai/dop/preview), Seedance Pro (bytedance/seedance/v1/pro/image-to-video), Kling 2.1 Pro (kling-video/v2.1/pro/image-to-video)

One tip: those IDs are what you'll see in the dropdowns on the pack's generation nodes, so this is also a handy reference when a model name in a guide doesn't match what you're seeing in your UI. The gap between the catalog and reality is a reminder that these are closed, hosted models - the whole point of the pack is that you're renting access to them per generation, not owning anything locally.