[Legacy] H3 Continuum Assemble

H3 Continuum Assemble is the V1-era splice node, and it's worth being precise about what it is not. It's not the V3 assembler (H3 Continuum Assemble V3.2.4 or the current Assemble + Seam) that reads an assembly_plan and manages chunk lists. This one is simpler: it appends two already-trimmed Continuum segments - "previous" and "next" - into one output while preserving exact audio sample order. That's the whole contract, and the description says it plainly.

Inputs: previous_images, previous_audio, next_images, next_audio. Outputs: images, audio (the concatenated result), and report (STRING). No options, no seam correction, no duration math - the trimming and alignment are expected to have happened upstream in Finish.

Where it lives in the old graph: your manual chain generates segment after segment via Join → sample → Finish, and Assemble progressively appends each finished segment to the accumulating result. It's the "concat" step of the V1 loop, and its exact-audio-sample-order guarantee is the thing that keeps the soundtrack from drifting across a long build.

Install

cd ComfyUI/custom_nodes
git clone https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum.git

or via ComfyUI Manager ("H3 Continuum"), then restart. No extra pip packages; needs a MiniMax H3 ComfyUI environment (weights not bundled) and ComfyUI v0.32.0+.

The trap with this node is feeding it untrimmed segments. If the "next" segment still carries its continuation overlap, you'll get duplicated frames at the splice and audio that stutters - this node assumes Finish already handled that. And don't reach for it in a modern workflow: H3 Continuum Assemble + Seam replaces it entirely, taking the sampler's assembly_plan, handling the overlap math itself, and adding the guarded video/audio seam correction. Assemble is the node you'll mostly see inside V1-era saved workflows and tutorials explaining the pack's history.