ℹ️ Flux LayerDiffuse Info

Some nodes do work; this one does the README's job for you. FluxLayerDiffuseInfo is a single-action info node with one input - action - and one output: an info string. It's the pack author's way of keeping the "what models do I need and where do they go" checklist inside the graph where you can actually see it while you're stuck.

The two actions

check_files (default) - scans your model directories and reports what's present and what's missing for the Flux LayerDiffuse pipeline.
setup_guide - prints the setup instructions and download links.

Run one, read the output, wire nothing else. It's not an output node in the blocking sense - you can drop it anywhere in a workflow or just double-click it in a blank graph and read the string.

What it's checking for

The LayerDiffuse-on-Flux stack needs a specific set of files, and they're spread across four ComfyUI model folders:

flux1-dev-fp8.safetensors → models/diffusion_models/
layerlora.safetensors → models/loras/ (the README says this one ships inside the pack, so it should already be there)
TransparentVAE.pth → models/vae/ (also bundled)
clip_l.safetensors + t5xxl_fp8_e4m3fn.safetensors → models/clip/
ae.safetensors → models/vae/

Download sources in the README: the Flux repo and RedAIGC's Flux-version-LayerDiffuse.

Why you'd bother

If you've installed this pack and are staring at "TransparentVAE.pth not found" or a loader that refuses to start, this node is the fastest way to see the whole file matrix at once instead of poking through folders. It's also genuinely useful as a mental checklist for the one thing that trips everyone: the LayerDiffuse pipeline is four separate components (base Flux, the layer LoRA, the transparent VAE, and the CLIP stack), and each has its own failure mode if missing. The setup_guide output is effectively the README's quick workflow rendered as a to-do list.

Honest take

It's a documentation node, so the quality ceiling is "helpful README in a box." Don't expect it to fix anything - it reports. If check_files says everything's present and you still get a tensor error, that's the CLIP-type problem (DualCLIPLoader type="flux", Conditioning Inspector, Conditioning Fix), not a missing file, and this node can't see that. And as with the rest of this pack: it's a one-author, near-zero-community footprint project, so this built-in guide is likely the best help you'll get. The pack's own README and the GitHub issues page are the rest of it.

Install

ComfyUI Manager → "DenRakEiw Nodes", or git clone https://github.com/DenRakEiw/DenRakEiw_Nodes, pip install -r requirements.txt, restart.