Nodes/IAMCCS-nodes/WanImageMotionPro Plus Simple
ComfyUI Node

WanImageMotionPro Plus Simple

The motion Pro with the scary inputs stripped off

By IAMCCS·Created 11 months ago·Updated 7 days ago· 113
WanImageMotionPro Plus Simple
  • positive
  • negative
  • anchor_samples
  • prev_samples
  • end_samples
  • clip_vision_start_image
  • clip_vision_end_image
  • positive
  • negative
  • latent
  • trim_slots
length81
motion_latent_count1
motion1.15
motion_modemotion_only (prev_samples)
add_reference_latentsfalse
vram_profilenormal
include_padding_in_motionfalse
safety_presetsafe
use_end_frametrue
end_lock_slots1
lock_start_slots1
end_overshoot_slots0
latent_refresh0.00
delta_max0.0
clip_vision_modeauto
prev_skip_last_slots0
continuity_profileoff
preset[custom]

The naming is doing a lot of work here, but it's accurate: IAMCCS_WanImageMotionProPlus_Simple ("WanImageMotionPro Plus Simple") is the same motion/end-lock engine as the Plus node with the intimidating internals taken out. No raw_mode, no prev_tail_pick_mode, no prev_anchor_blend, no prev_samples_profile, no diagnostic logging - just the motion controls, the end-frame lock, and the presets. If you want the FLF + motion behavior without a PhD in latent-tailing, this is the one to start with.

Everything that makes the family useful is still here. motion (1.15 default) boosts the prev-sample tail to kill SVI Pro's fake slow-motion. use_end_frame + end_lock_slots lock the final slots to end_samples. end_overshoot_slots extends the window so the clip converges to the end frame instead of freezing, and you trim the excess with trim_slots. clip_vision_mode is here too, including end_only - the fix for crossfade bleed when your anchor and end frame are the same image, which is most of SVI Pro continuation.

What got simplified away

The tail-handling deep-dive is what was cut: on the full node you can pick which part of the prev tail counts as motion, skip the last N slots, blend the tail toward the anchor, and choose aggressive vs. safe SVI profiles. The Simple version keeps just prev_skip_last_slots - "ignore the last N tail slots if the carried SVI tail is too frozen" - and calls it a day. That's a reasonable trade. For the common case where your prev tail is well-behaved, you don't miss the rest.

What survived that matters: latent_refresh and delta_max (the AdaIN stabilization for long chains), safety_preset (base/safe/safer/legacy), vram_profile, lock_start_slots, and the preset dropdown with the same server-side recipes as the full node.

Outputs: positive, negative, latent, trim_slots - identical contract to the rest of the family, so you can swap this node in and out of a workflow without rewiring.

Install

Standard IAMCCS-nodes:

cd ComfyUI/custom_nodes
git clone https://github.com/IAMCCS/IAMCCS-nodes.git

or ComfyUI Manager → search "IAMCCS" → restart. No models, no extra dependencies. ComfyUI ≥ 0.3.0, Python ≥ 3.12, Torch ≥ 2.8.

Caveats

The family warnings still apply: keep end_transition_frames-style blending away (this node doesn't even expose it, which is the point), and watch the KJNodes FETA/wrapped-attention crash if you run end_overshoot_slots > 0 - set it to 0 or update KJNodes. The honest verdict: for most SVI Pro users this Simple variant is the right default, and you only graduate to the full Plus node when you hit a chain that genuinely needs custom tail handling.

CategoryIAMCCS/Wan

Inputs (25)

NameTypeDefaultDescription
positiveCONDITIONING
negativeCONDITIONING
lengthINT811–16384
anchor_samplesLATENT
motion_latent_countINT10–16
motionFLOAT1.151–2
motion_modeCOMBOmotion_only (prev_samples)2 options: motion_only (prev_samples), all_nonfirst (anchor+motion)
add_reference_latentsBOOLEANfalse
vram_profileCOMBOnormal5 options: normal, chunked_blocks_2, chunked_blocks_4, loop_per_frame (lowest_vram), cpu_offload (slowest)
include_padding_in_motionBOOLEANfalse
safety_presetCOMBOsafe4 options: base, safe, safer, legacy
use_end_frameBOOLEANtrue
end_lock_slotsINT10–16
lock_start_slotsINT10–16
end_overshoot_slotsINT00–8
latent_refreshFLOAT0.000–1
delta_maxFLOAT0.00–5
clip_vision_modeCOMBOauto4 options: auto, end_only, start_only, none
prev_skip_last_slotsINT00–16Ignore the last N temporal latent slots from prev_samples before extracting the motion tail. Useful when the last part of a carried SVI tail is too rigid, frozen, or over-constrained.
continuity_profileCOMBOoff8 options: off, SVI_CONTINUITY, SVI_CONTINUITY_PREV_SAFE, SVI_CONTINUITY_PREV_AGGRESSIVE, SVI_CONTINUITY_END_ONLY, SVI_CONTINUITY_MOTION_SAFE, +2
presetCOMBO[custom]13 options: [custom], raw, raw overshoot, svi continuity, svi continuity prev safe, svi continuity prev aggressive, +7
prev_samplesoptLATENT
end_samplesoptLATENT
clip_vision_start_imageoptCLIP_VISION_OUTPUT
clip_vision_end_imageoptCLIP_VISION_OUTPUT

Outputs (4)

NameTypeDescription
positiveCONDITIONING
negativeCONDITIONING
latentLATENT
trim_slotsINT