Checkpoint Loader (Simple)

At a glance this looks like the plainest node in the pack, and mostly it is - it loads a Stable Diffusion checkpoint and gives you back the MODEL, CLIP and VAE, same as core ComfyUI's own loader. The one thing it adds is the reason WAS made it: it also outputs the checkpoint's name as a text string. The README describes the whole WAS loader set the same way - "loaders that also output a string that contains the name of the model being loaded."

That extra string is small but genuinely useful. Wire it into a Save Image path or filename and every output gets stamped with which checkpoint made it. No more staring at a folder of PNGs six months later trying to remember whether that batch was DreamShaper or Juggernaut.

How it works

Functionally it's the standard checkpoint load: you pick a .safetensors (or .ckpt) from your ComfyUI/models/checkpoints/ folder, and the node splits it into the three things the rest of a text-to-image graph needs - the diffusion MODEL, the CLIP text encoder, and the VAE for decoding latents to pixels. The WAS twist is a fourth output carrying the filename you selected, as a plain string, so downstream nodes can read it.

Nothing about the loading itself is special or faster than core - this isn't an optimization, it's a convenience. If you never need the name string, core's Load Checkpoint does the identical job. If you're building a workflow that labels its own outputs or branches on which model is loaded, this saves you a separate lookup.

Inputs and outputs that matter

There's really one input that matters: the checkpoint name dropdown, populated from your checkpoints folder. Pick your model and you're done.

The outputs are the interesting part:

• MODEL → your KSampler.
• CLIP → your text encoders (or WAS's own Text to Conditioning).
• VAE → your VAE Decode.
• the name string → a Save Image filename, a text overlay, or anything that wants to record what produced the image.

How to install it

Via ComfyUI Manager: search was-node-suite-comfyui, install, restart. Or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/WASasquatch/was-node-suite-comfyui

then run the pack's requirements.txt in your ComfyUI venv and restart. The loader itself needs no extra downloads - it reads checkpoints you already have.

Common issues & troubleshooting

Your checkpoint doesn't appear in the dropdown. It has to live in ComfyUI/models/checkpoints/ (or a path mapped in extra_model_paths.yaml). Drop the file in and restart, or hit refresh - new files don't always show up live.

The VAE output looks washed out or wrong. Some checkpoints ship a baked-in VAE and some don't. If colors look off, don't blame the loader - load a dedicated VAE separately and feed that to your decode instead of the one that came out of the checkpoint.

Why use this over the built-in loader? Only for the name string. If you don't need it, core's Load Checkpoint is one less custom-node dependency, which matters because this pack is heavy. Don't pull in all of WAS Node Suite just for a checkpoint loader - reach for it when you're already using the suite's text and save nodes and want them to know what model ran.

"Import Failed" for the whole suite after a ComfyUI update. The standard WAS gotcha: the pack downgrades some packages for its older modules, and a ComfyUI upgrade can break the import at startup. Re-install the pack's requirements against the activated venv (or run install.bat). The suite has been in maintenance mode since late 2023, so this is a known cost of keeping it around.