📥 S42 CutFlow Frame Insert

S42CF_FrameInsert is the counterpart to FrameExtract and FrameHold - where those pull frames out or stretch them, this one splices new frames into a clip at a chosen position. It's the node you reach for when you want to prepend a title card, patch a corrected sequence back into a video, or stick a freeze-frame in the middle of a segment.

What you set

Three inputs, and the tooltips carry the load:

• clip - the main IMAGE batch you're inserting into.
• insert_frames - the IMAGE batch to splice in. The important detail: it gets auto-resolution-matched to the main clip. You can insert a 512×512 title card into a 1920×1080 video and it'll be scaled to fit - no manual resizing before you wire it up.
• insert_at - the frame index where insertion begins. 0 prepends (your new frames become the new head of the clip). The tooltip's hint: use the main clip's frame_count to append at the end. Anything between inserts mid-clip.

Outputs are clip, frame_count (now the sum of both lengths), and info.

The obvious workflow: hold + insert = pause button

FrameInsert and FrameHold are natural partners. The classic move:

1. S42CF_FrameHold on your clip at custom / frame N → produces N duplicated frames.
2. S42CF_FrameInsert takes those frozen frames and splices them into the original clip at the same spot.

Net result: a genuine pause with zero re-encoding or external editing. You can also use FrameExtract to pull a handful of frames, process them separately (inpaint a glitch, color correct one region), then Insert them back where they came from - the patch-and-return pattern that's shockingly handy when one section of an otherwise good render is broken.

Also worth noting: because insertion is just tensor concatenation on CPU, it's instant. No re-render, no model, no VRAM. And since every CutFlow node speaks standard IMAGE batches, the insert_frames input can come from anywhere - another clip, an ImageToClip still, a generated batch.

Install

Same one-pack story. ComfyUI Manager → search "S42 CutFlow" → Install → restart ComfyUI. Manual:

cd ComfyUI/custom_nodes/
git clone https://github.com/GeekyGhost/S42-CutFlow.git
pip install -r S42-CutFlow/requirements.txt

Restart fully; the console should print [S42 CutFlow] Loaded 53 total nodes. Pure CPU, no downloads beyond the pack. Only required dependency is OpenCV (opencv-python-headless in the shipped requirements.txt; README says pip install opencv-python if you see "No module named 'cv2'").

Pack health note, once more, because it's true for every node here: Geeky Ghost ships S42 CutFlow as a work-in-progress being refined for LTX Desktop. Clip-ops like this are the solid core; the expanded nodes are where the churn lives.