IAMCCS Cine FLF Productor Dyno

Cine FLF Productor Dyno is the Cine FLF Productor plus a single extra control - and that control is the interesting part. The Dyno family in this pack layers legacy guide-attention entries onto the guide-latent mechanism; this node is the plan-driven production version of that idea, distinguished by one enum: guide_attention_mode. It's the node you pick when you want the full guide plan and the ability to toggle dynamic guide attention without changing anything else in your graph.

How it works

Everything about the base Productor holds: guide_plan_json drives the schedule (falling back to timeline_data), multi_input supplies the reference images, vae encodes them, and strength_scale/tail_safety_frames tune the result. On top of that, Dyno builds guide-attention entries - the conditioning-side reinforcement that keeps motion-heavy shots locked to their references. The guide_attention_mode switch decides how those entries are generated:

• skip_step_transitions (default) - the author's recommended production mode: keeps the step-transition pairs off the clean guide path, which the tooltip explains as "keeps step-transition pairs on the clean Production guide path." In other words, you get the attention reinforcement where it helps without polluting the transition handling.
• all_guides - add attention entries for every guide; maximum reinforcement, more cost, and can stiffen motion.
• off - Dyno behaves like the plain Productor; useful as an A/B switch when you suspect the attention pass is hurting a shot.

The same graceful-degradation caveat from the Simple Dyno engine applies: the attention helper is imported from ComfyUI internals with a try/except, so on an older ComfyUI the attention pass silently doesn't happen. Keep ComfyUI current if you rely on it.

Installing it

Part of IAMCCS-nodes:

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

Restart ComfyUI, or search "IAMCCS" in ComfyUI Manager. Baseline: ComfyUI ≥ 0.3.0, Python ≥ 3.12, Torch ≥ 2.8.

The honest take

The guide_attention_mode enum is more useful than it looks, because "off" gives you a one-widget A/B test of the whole Dyno feature - leave the graph intact and flip between off and skip_step_transitions to see whether the attention pass is helping or hurting a given take. If you're not seeing a difference, check your ComfyUI version first, because the silent-fallback path means you might be comparing Dyno to Dyno. Start on the default, and only drop to all_guides for genuinely chaotic motion. This is the right node when your shots are plan-driven and motion-heavy - the one case where the plain Productor often isn't enough.