IAMCCS WAN Load Image From Board PURE

In IAMCCS's shotboard pipeline, the "board" is the plan: which images open each shot, which close it, at what resolution, resized how. IAMCCS_WanLoadImageFromBoard ("IAMCCS WAN Load Image From Board PURE") is the node that turns that plan into an actual IMAGE tensor without you hunting down file paths. You give it the wan_shotboard and a role - first, middle, or last - and it finds the right image, loads it, and applies the board's own resize settings.

This is the "PURE" philosophy in action: the board carries the resolution (image_width/image_height, default 832×480), the resize method (default crop), and the 16-pixel multiple, and this node just obeys. No per-workflow guessing about whether your frames need to be 16-aligned - the board already decided.

The inputs

• wan_shotboard - the IAMCCS_WAN_SHOTBOARD object (from a planner/board node). The source of truth for paths and sizing.
• role - which frame you want. first for the shot's opening anchor, last for the end-frame target, middle when you want something from the middle of the shot.
• debug_verbose - prints the resolved path and dimensions to the console when you're chasing a "wrong image loaded" bug.

Outputs are just image (a single-frame IMAGE tensor) and report (JSON with the role, resolved path, final dimensions, and resize settings).

Why you'd use it instead of a plain Load Image

Two reasons. First, consistency: the board's resize settings are applied automatically, so every chunk in a 40-shot sequence is guaranteed to come out at the same dimensions with the same crop behavior. Second, path handling: you don't type or wire a path at all - the shotboard resolves it, and if the file is missing, the report tells you the exact path it tried.

It's also a neat way to grab an end frame for FLF-style work: wire the shotboard, set role = last, and feed the result into your end-frame encoder while IAMCCS_WanFLFPairFromTimeline handles the paired start image.

Install

Same pack as everything else:

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

or ComfyUI Manager → search "IAMCCS" → restart. No model downloads, no new dependencies. Compatibility floor: ComfyUI ≥ 0.3.0, Python ≥ 3.12, Torch ≥ 2.8.

The catch

The middle role depends on the board actually defining a middle image - if your shotboard only stores first/last pairs (which is the FLF norm), middle may fall back or error depending on how the board was built. And since the resize comes from the board, don't be surprised if first loads at a different aspect than the raw source file; that's the crop/pad doing its job. If the wrong image shows up, flip debug_verbose on and read the path it resolved.