📊 Transparent Image Info

The pack's transparent-image trio ends with a checker. TransparentImageInfo takes an image, inspects its alpha channel, and returns an info string with statistics - and its quiet job is answering the one question that decides whether your LayerDiffuse run actually worked: did the pipeline produce real transparency, or did the alpha get flattened somewhere?

What it does

One input - images (IMAGE, required) - and one output - info (STRING). It analyzes the alpha channel of the incoming image and reports on what's actually there: whether an alpha channel is present at all, the amount/coverage of transparency (how much of the image is genuinely transparent vs. opaque), and related statistics. It's the diagnostic half of the Preview/Info pair: Preview shows you the transparency visually, Info puts a number on it.

That last part matters more than it sounds. The most common transparent-pipeline failure isn't a crash - it's a silent flattening where the alpha gets dropped (an RGB save, a decode that didn't use the transparent VAE, a graph that didn't preserve RGBA end-to-end) and the image comes out looking fine but fully opaque. That's the failure mode the layerdiffusion ecosystem's own knowledge base calls out: "not every ComfyUI node preserves the alpha channel through a pipeline, so some graphs silently drop it." TransparentImageInfo is the cheap canary for exactly that. Feed it the decoder's output, and if the info string says there's no alpha or near-zero transparency, you know the problem is upstream - usually the transparent VAE never actually engaged.

Where it fits

Drop it after FluxLayerDiffuseDecoderSimple, next to PreviewTransparentImage. If you're building a workflow you'll share, keeping this node in the graph means anyone who runs it can verify the transparency in one glance instead of trusting that "it probably worked." It also slots in as a guard: wire its info string into a text display or use it to decide whether to proceed with a save.

The honest take

It's a three-line tool - the whole pack treats it as a utility alongside Save and Preview - and that's fine. It won't fix a broken alpha, and it can't tell you whether your edges are good (that's what your eyes are for, on a checkerboard preview). What it does is turn "did the alpha survive" from a guess into a fact, which is the single most useful diagnostic in this whole transparency subsystem. Use it once to verify your chain, and keep it around for when you change something and need to re-verify.

Install

Part of DenRakEiw_Nodes - ComfyUI Manager → search "DenRakEiw Nodes", or:

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

then restart. Same one-author-pack caveat as its siblings: if the statistics ever look wrong, the GitHub issues page is your only support. For a diagnostic node that's an acceptable trade.