Get Sampler Params Extra

Get Sampler Params Extra is the richest member of the pack's extras getter family, because it reads back an entire sampling recipe - sampler, scheduler, steps, denoise, seed, and cfg - from a single stored value, and then splits it into all seven of its output sockets for you. One Set Sampler Params Extra upstream, and every knob your sampler needs comes out on its own wire.

Why bundle sampler settings at all

This pack's take on the standard ComfyUI sampler spaghetti is the sampler_params bundle: instead of dragging sampler, scheduler, steps, seed, and CFG around as loose values, you aggregate them into one object (the Sampler Params node does this), pass the bundle where it's needed, and split it when a node wants individual values. Get Sampler Params Extra is the version that retrieves such a bundle from image metadata. It's the answer to "I want this generation's exact sampling settings available next time, without rebuilding the recipe by hand."

Because it rides in image_info extras, the recipe survives a save/load cycle: Image Saver writes it into the file, Image Reader restores it, and the settings you tuned are right there again.

Inputs and outputs

Inputs: the family-standard extras socket and a key for lookup.

Outputs - this is where it gets fun:

• sampler_params - the bundle itself, for pack nodes that accept it whole.
• sampler and scheduler - the names, ready for KSampler-style sockets.
• steps (INT), denoise (FLOAT), seed (INT), cfg (FLOAT) - the numbers, each on its own typed socket.

If the key is missing you get Nones across the board, which will make a sampler very unhappy - this is a "lookup, not guarantee" node like its siblings, so keep the writer connected.

A note on those numbers

A stored recipe is a promise you made to yourself last week. Sampler settings that worked at 20 steps might look "deep fried" if you swap them onto a distilled or Turbo-style model, where the community's troubleshooting wisdom is that CFG and step counts behave completely differently. Reusing settings is the point; reusing them blindly is how you get oversaturated output. Keep the recipe, sanity-check the model.

Installing it

Part of the ComfyUI-Info-Prompt-Toolkit pack. ComfyUI Manager - search "Info Prompt Toolkit" - or:

cd ComfyUI/custom_nodes
git clone https://github.com/kinorax/comfyui-info-prompt-toolkit.git
cd comfyui-info-prompt-toolkit
pip install -r requirements.txt

Restart ComfyUI. No models to download. Small single-author pack (GPL-3.0) with a modest footprint - the bundle-then-split pattern is its take on taming sampler wiring.