Slider Receiver

Slider Receiver is the boring half of the Dynamic Sliders Stack pair, and it's supposed to be. The whole pack exists so you can keep a wall of sliders on your canvas without dragging a hundred wires across your graph: the Slider Stacker squashes every slider value into one comma-separated string, and this node unpacks it. You feed it that string, tell it which slider you want by number, and it hands you a single FLOAT. That's the entire job. No API, no model files, no hidden behavior - it's a tiny pluck node, and the less you have to think about it, the better it's working.

Where it earns its keep is in the wiring. Say you've got a Stacker with five sliders controlling prompt weights for a prompt-composer node, or five LoRA strengths. Without this pack you'd either eyeball a bunch of separate number boxes or build something far more convoluted to batch-adjust them. With it, one string wire leaves the Stacker, and as many Receivers as you need - one per slider - tap off it and hand clean floats to whatever wants them. Add more sliders to the Stacker later and nothing downstream moves; you only add Receivers. That separation is the whole design, and it's why the pack's v2.0.0 API split the pair this way.

Mechanically it's as simple as the pitch. The receiver splits weights_string on commas, trims whitespace, and reads the value at index (1-based). Two gotchas live in that implementation. First: it's 1-based, so slider one is index 1, not 0. Second - the one that will actually bite you - an out-of-range or invalid index doesn't error out; it quietly returns 0.0 and prints a warning to the console. A slider that reads zero isn't necessarily set to zero, it may just be pointing past the end of the list. If a weight "stops working", check your index against the Stacker's slider_count before you blame the node.

The inputs you actually set:

• weights_string - connect the SLIDER_WEIGHTS output from the Slider Stacker here. The author's own description is blunt: that's where the stacker's string goes.
• index - which slider (1–50) to pull out. Default 1.

The single output is WEIGHT, a plain FLOAT, and it'll wire into anything that takes a float: prompt composer strength, a LoRA loader's strength, denoise, CFG, whatever. That's the whole point of the node - it turns a crowded string into a number the rest of the graph understands natively.

Install is the standard two-liner. Use ComfyUI Manager (search "Dynamic Sliders Stack", or paste the repo URL into the install field) or clone it by hand:

cd ComfyUI/custom_nodes
git clone https://github.com/FunnyFinger/Dynamic_Sliders_stack.git

Then restart ComfyUI and look under the "Dynamic Sliders Stack" category - the node shows up as Slider Receiver (the "(DSS)" is only its internal registration name). Good news on weight: pyproject.toml lists zero dependencies and there are no model files to fetch, so install is genuinely painless and there's nothing extra to download.

One honest caveat about the use case, not the node. A Stacker-and-Receiver rig feeding prompt weights only does anything on models whose encoders honor attention weights - SD 1.5, SDXL and its descendants like Illustrious and Pony. On Flux and the LLM-encoder models, (word:1.3)-style weighting is discarded by the text encoder, so sliders feeding a composer on those checkpoints are moving a dial that isn't connected. The pack is great; just know which pipeline you're aiming it at. Pair it with a prompt composer and a LoRA loader on SDXL-lineage models and it's a tidy little control room.