Anchors every guide of a MiniMax H3 shot from one node, instead of a row of Add Guide for MiniMax H3 nodes each dragging the same four cables in from far upstream. Move one and the canvas turns to spaghetti.
flowchart LR
POS(["positive"]):::input --> MG
VV(["video vae"]):::input --> MG
AA(["audio vae"]):::input --> MG
LAT(["latent"]):::input --> MG
IMG(["image"]):::input -- guide 1 --> MG
VID(["video"]):::input -- guide 2 --> MG
AUD(["audio"]):::input -- guide 3 --> MG
MG["**NKD MiniMax Guides**"]:::nkd --> O1(["positive"]):::output
MG --> O2(["video vae"]):::output
MG --> O3(["audio vae"]):::output
MG --> O4(["latent"]):::output
classDef nkd fill:#3b3b6b,stroke:#8ab4ff,stroke-width:2px,color:#fff
classDef input fill:#2d2d2d,stroke:#888,color:#eee
classDef output fill:#1f4a1f,stroke:#7fd97f,color:#fff
- The guide list grows as you fill it. Each slot takes a still, a frame sequence, a video (picture and its soundtrack) or bare audio.
- A
positionwidget appears next to each filled slot, for the frame it lands on. Negative positions count from the end. - Two slots on the same position, a clip and its sound, become one guide.
latent,video vaeandaudio vaecome straight back out, so the sampler and whatever comes next hang off this node instead of reaching back across the graph. Both VAEs are required inputs rather than optional ones, which keeps them at the top of the node, above the guide list that grows.
The anchoring itself is the core node's: same clip-length snapping, same audio cropping, same errors.