IAMCCS WAN Shotboard Loop Info

The IAMCCS_WanShotboardPlannerPure node hands you an opaque IAMCCS_WAN_SHOTBOARD bundle. This is the node that opens it up. IAMCCS WAN Shotboard Loop Info takes that bundle and unpacks it into the numbers you can actually wire into the rest of your graph.

What it does

It takes exactly one input, cine_linx (the shotboard from Planner PURE), and returns five things:

• timeline_plan - the IAMCCS_WAN_TIMELINE_PLAN, the narrowed-down plan object that the loop-select nodes expect. This is the output you'd pass forward into the loop.
• chunk_count - how many FLF start→end chunks the board has. Wire this into your loop counter's upper bound so you know when to stop.
• total_frames - the whole board's frame count at the planned frame rate.
• frame_rate - the FPS the plan was built at.
• report - a JSON string with the full breakdown: schema, chunk count, total frames, how many timeline images, FLF pairs, and prompt-relay chunks the plan carries. Handy for debugging, and it doubles as a "did the planner see what I think it saw" check.

That's the whole job. It's a read-only diagnostic-and-routing node, not a processing stage - nothing gets modified, no models touched, no latents harmed.

Why you'd bother

Because the shotboard bundle is deliberately a black box. When a long loop misbehaves three iterations in, the first question is "what did the planner actually decide?" - and this node is the answer without you having to dump JSON into a debug text node. It also gives you chunk_count and total_frames as clean INT outputs, which is exactly what a primitive loop needs to know when to stop and how long the final video should be.

Where it fits

Planner PURE → Loop Info → Loop Chunk Select → motion node → sampler → (feed latents back as prev_samples) → next iteration. Loop Info is the stage that hands the loop its marching orders: the timeline_plan it passes out is the same plan IAMCCS_WanShotboardLoopChunkSelect will slice up per iteration.

Installing it

It ships in IAMCCS-nodes. ComfyUI Manager (search "IAMCCS") or:

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

Restart after. No extra dependencies, no model downloads - this node is pure plan-math and JSON. Same house rule as the rest of the pack: keep it updated, because CCS ships versions fast and saved workflows can silently reference renamed nodes.

Common issues

If chunk_count comes back as 0 or 1 when you expected a multi-shot board, the problem is upstream in Planner PURE - you need at least two image slots to form an FLF pair. Loop Info is just reporting what the planner built; don't blame the messenger.