Nodes/AnimateDiff Evolved/Set Props on Conds πŸŽ­πŸ…πŸ…“
ComfyUI Node Runs on cloud

Set Props on Conds πŸŽ­πŸ…πŸ…“

Masking a positive/negative pair together

By KosinkadinkΒ·Created 3 years agoΒ·Updated 8 days agoΒ· 3,519
Set Props on Conds πŸŽ­πŸ…πŸ…“
  • positive_ADD
  • negative_ADD
  • opt_mask
  • opt_lora_hook
  • opt_timesteps
  • positive
  • negative
β—„strength1.00β–Ί
β—„set_cond_areaβ–Ύβ–Ί

Regional prompting in AnimateDiff means telling part of your conditioning "you only apply here, this strongly, maybe only during this stretch of sampling." This pack has a single-conditioning family that does that one branch at a time (Set Props on Cond - singular). This node does the same job for people working in positive/negative pairs the whole way through: mask, strength, LoRA hook, and timestep range, applied to a positive and a negative branch together in one node instead of two.

What it does

positive_ADD/negative_ADD are the pair you're modifying. strength (default 1, range 0–10) scales how strongly this pair pulls once combined with whatever else is in your graph - leave it at 1 unless you have a specific reason to over- or under-weight this region. set_cond_area is the same option ComfyUI's native masked-conditioning nodes use: broadly, whether the mask's coverage is read precisely or tightened toward a bounding-box interpretation of where it's actually nonzero.

The three optional inputs are what make this an AnimateDiff-Evolved node rather than a stock ComfyUI one. opt_mask is the actual spatial mask - without one, you're just re-setting strength on the whole pair, not masking anything. opt_lora_hook scopes a LoRA hook (from the pack's LoRA Hooks system) to fire only within this masked, conditioned region rather than globally. opt_timesteps scopes the same pair to a range of the denoising process itself, so it's active for only part of the sampling steps rather than the whole render.

The inputs and outputs that matter

Required: positive_ADD, negative_ADD (CONDITIONING), strength (default 1, range 0–10), set_cond_area.

Optional: opt_mask (MASK), opt_lora_hook (HOOKS), opt_timesteps (TIMESTEPS_RANGE).

Outputs: positive, negative - ready to feed Manual Combine Conds alongside other branches, or straight into your sampler if this is your only one.

How to install it

ComfyUI Manager: search AnimateDiff Evolved, confirm the author is Kosinkadink, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/Kosinkadink/ComfyUI-AnimateDiff-Evolved

then restart. No model download tied to this node itself - the motion module your workflow needs elsewhere (mm_sd_v15_v2 and friends) is the actual download requirement.

Common issues & troubleshooting

Not seeing any regional effect at all. A mask-less call to this node isn't spatial - it just re-sets strength on the whole pair. Confirm opt_mask is actually wired in before assuming the node is broken.

Strength above 1 seems to let this region dominate everything else. The 0–10 range isn't a normal 0–1 blend weight - it doesn't clamp you, and pushing it well past 1 tends to let this region overwhelm whatever it's combined with later. Start at 1 and nudge up in small steps.

LoRA hook attached here doesn't seem to fire. opt_lora_hook needs a properly built hook from the pack's LoRA Hooks system (registered and, if you're scheduling it, given keyframes) - a raw LoRA loader output isn't the right type here.

Not sure whether you want this or the singular version. If your graph builds conditioning as one branch at a time rather than positive/negative pairs, the sibling Set Props on Cond (no "s") does the equivalent job for that family instead.

CategoryAnimate Diff πŸŽ­πŸ…πŸ…“/conditioning

Inputs (7)

NameTypeDefaultDescription
positive_ADDCONDITIONINGβ€”
negative_ADDCONDITIONINGβ€”
strengthFLOAT1.000–10β€”
set_cond_areaCOMBO2 options: default, mask bounds
opt_maskoptMASKβ€”
opt_lora_hookoptHOOKSβ€”
opt_timestepsoptTIMESTEPS_RANGEβ€”

Outputs (2)

NameTypeDescription
positiveCONDITIONINGβ€”
negativeCONDITIONINGβ€”