H3RelayInternalChainContext

Apply each scene's inherited or overridden guide, sliding history, or masked AV continuation, including scene 1 when Existing Video Context is connected.

By akatz-ai·Created 8 days ago·Updated 4 days ago· 9

state
conditioning
vae
latent
audio_vae

conditioning
trim_frames
is_continuation
latent

Categoryconditioning/minimax/contex_loop

Inputs (5)

Name	Type	Description
state	H3_CHAIN_STATE	Current state from H3 Chain Current Shot.
conditioning	CONDITIONING	Conditioning from the stock MiniMax H3 Ref2VA/I2V node. Scene 1 passes through without motion context; later scenes receive the saved continuation context.
vae	VAE	MiniMax H3 video VAE used to encode saved context frames for continuation scenes.
latent	LATENT	The CURRENT scene's empty AV latent from the stock H3 conditioning node. Chain Context passes it through in guide mode or returns a masked preserved-prefix copy in masked_av mode.
audio_vaeopt	VAE	H3 audio VAE used only when scene 1 continues from imported video audio. Later loop scenes reuse their saved AV latent directly. It may be left disconnected for visual-only context or source_track mode.

Outputs (4)

Name	Type	Description
conditioning	CONDITIONING	Conditioning ready for the H3 guider/sampler: scene 1 passes through unless Existing Video Context seeds it; later scenes always continue.
trim_frames	INT	Repeated leading frames to remove after decoding. Connect to MiniMax H3 Contex Loop Trim.
is_continuation	BOOLEAN	True for resumed/continued scenes, false for the first scene.
latent	LATENT	Sampler-ready target latent. In guide mode this is the input latent unchanged. In masked_av mode its preserved AV prefix and nested denoise mask carry the previous scene into the current target. Wire this output to the sampler for both modes so Plan can switch safely.