Get Rating Class From Text

Unlike most of this pack's tagger nodes, this one genuinely does something different from what its plain-named sibling (GetRatingNode) does. Where GetRatingNode wants a live IMAGE tensor from inside your ComfyUI graph, this node's image input is a plain STRING - a filesystem path, defaulting to a placeholder like /path/to/image.jpg. It opens the file at that path itself rather than expecting you to have loaded it into the graph already.

That makes it the right tool for a different kind of workflow: batch-processing files that already exist on disk rather than images you just generated. Feed it paths from a directory-listing node, a loop over a folder of existing images, or a dataset manifest, and it runs the same content-rating classification GetRatingNode runs in-graph - but without needing a Load Image step for every single file first.

How it works

model_name (default EVA02_Large) picks the tagger backbone - same option set as the rest of this pack's tagger nodes: ViT_Large, SwinV2, ConvNext/ConvNextV2, MOAT, ViT, and _v3 revisions of a few of those. The node reads the file at the given path, runs it through that model, and returns a rating class as a string, following the Danbooru rating vocabulary these models were trained on (general/sensitive/questionable/explicit, roughly - the exact label set depends on the underlying model version).

Inputs and outputs

• image (default /path/to/image.jpg) - a filesystem path, as a STRING, not an IMAGE tensor. This is the detail that trips people up coming from the rest of the pack.
• model_name (default EVA02_Large) - the tagger backbone.

Output is a single STRING - the rating class.

Installing ComfyUI-LogicUtils

ComfyUI Manager: Install Custom Nodes → search "ComfyUI-LogicUtils" → install → restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/aria1th/ComfyUI-LogicUtils

Restart ComfyUI. Same as the rest of the tagger family - expect a model download on first use of a given model_name, and if it fails with a missing-package error, check LogicUtils' auto-install hook (COMFYUI_LOGICUTILS_AUTO_INSTALL=1) or its requirements.txt before assuming something's broken.

Worth knowing

Double-check the path is reachable from wherever the ComfyUI process is actually running, not just from your own machine - if ComfyUI's running inside a container or on a remote box, /path/to/image.jpg needs to resolve on that filesystem. This is part of aria1th's (AngelBottomless's) LogicUtils pack - the same handle behind the Danbooru-trained Illustrious XL series, which is presumably why this repo carries a full WD-style tagger family in the first place, though LogicUtils itself is a separate, much smaller project. The README stops at "documentation is being prepared," and public commits ended in January 2026, so there's no official word on exactly which rating labels this returns beyond what the underlying model defines - good for triage, not a substitute for a real moderation policy.