8 Floats from List

8 Floats from List is the full unpacker for Mira's float bundles. Feed it a FLOAT_LIST and it fans the values back out as eight separate FLOAT outputs. It's the natural other end of the 8 Floats packer: gather a set of numbers into one wire on one side of the graph, run that single link across, and explode it back into individual values wherever they're used.

It's a Numeral node in ComfyUI_Mira, mirabarukaso's utility pack (there's also a single-output version for when you only need one value). The whole point of the packer/unpacker pair is to keep busy workflows readable - instead of eight float wires snaking across the canvas, you get one, with the fan-out happening only at the point of consumption.

How it works

Give it the list and a Start_At_Index, and it reads eight consecutive entries from that index, exposing them as float_1 through float_8. With the index at 0 you unpack a typical 8-value bundle in full; raise it and the window of eight slides further into the list. It's pure lookup - no math, no scaling.

The clean round-trip is the common shape: 8 Floats on the left, one wire across, 8 Floats from List on the right, and everything in between stays uncluttered.

Inputs and outputs that matter

• float_list - the FLOAT_LIST to unpack, usually from 8 Floats.
• Start_At_Index - where to start reading; the node takes eight from there. Zero-based.

Outputs are float_1 through float_8, eight individual FLOATs. Route each into whatever numeric input it belongs to - strengths, scales, blur amounts, sampler settings.

Installing it

ComfyUI Manager: search ComfyUI_Mira in the Custom Nodes Manager, Install, restart. Manual: cd ComfyUI/custom_nodes && git clone https://github.com/mirabarukaso/ComfyUI_Mira.git, then restart ComfyUI. If the pack errors on load, run pip install -r requirements.txt in its folder and restart. No models to download.

Common issues

This node always wants eight entries, so the pitfall is asking for a window the list can't fill - a non-zero Start_At_Index on an 8-element list runs off the end and some outputs have nothing to read. Keep the source list long enough for where you begin.

Two small ones: the outputs are labeled from float_1 but indexing is zero-based, so Start_At_Index 0 puts the list's first element on float_1 - mind the offset when a value lands one slot from where you expected. And if you only need a couple of values, the single-output interpreter is less clutter; use this one when you actually want the whole bundle back out. It's plumbing, and it does its plumbing without surprises as long as your indices stay in range.