Chroma Strip Overlay

There's a compositing trick that predates AI: put your overlay on a clean green strip beside the main footage, key the strip out in your NLE, and you've got a screen-insertable element with perfect edges - because your editor's chroma keyer, not a segmentation model, decides the matte. Chroma Strip Overlay is that trick as a node. It takes video frames, places a chroma-key green strip on one side, composites an overlay image inside the strip, and hands you a frame ready to key in Premiere Pro or Resolve. The tooltip says it plainly: "Designed for keying in NLEs like Premiere Pro."

Why would you want this instead of just generating alpha directly? Because NLE chroma keying is fast, keyframe-able, and lets the editor tune the matte interactively after the fact. If you're delivering a green-screen element for an editor who wants control - or you're building templates where the overlay changes but the pipeline doesn't - a clean green strip is the most editor-friendly delivery format that exists. The node is doing the deterministic prep so the editor's keyer has the cleanest possible material.

The layout controls

• side - left or right. Which edge carries the strip.
• mode - overlay (the strip covers the video edge) or expand (the strip adds extra width to the frame). Expand is the one to reach for when you don't want to lose any of the main footage width.
• strip_width_pct - strip width as a percentage of total frame width (default 25, range 5–50).
• image_padding_pct - padding around the overlay inside the strip, as a percentage of strip width (default 5).
• vertical_position - 0.0 = top, 0.5 = center (default), 1.0 = bottom. Where the overlay sits in the strip.

The green is a fixed chroma-key green (in the source, a literal float-RGB constant), which is what you want - the whole point is a consistent, ultra-keyable color. Output is a single image batch.

Where it slots in

The realistic pipeline: generate your overlay element (a logo, a caption card, a talking-head insert), place it on the strip, export, then key in the NLE. Since it operates on a frame batch, the same node renders the whole video with the strip on every frame - identical position, no per-frame drift, which is exactly what a keyer wants. And because the strip is added beside the footage rather than composited into it, the editor can swap, move, or scale the overlay without touching your ComfyUI graph again.

It pairs thematically with the pack's CorridorKey Green Screen Keyer - but that's the AI route (neural matting from a green screen). This node is the opposite direction: it manufactures a green screen for a traditional keyer. If you have both, you have every edge case covered: generate the insert as a strip here, or key real footage with CorridorKey there.

Gotchas

• mode: overlay covers part of the main video. If you can't afford losing frame width, expand is your mode - just know it changes the output aspect ratio, so downstream nodes that expect the original dimensions will complain.
• The overlay is scaled to fit the strip. If the strip is thin and the overlay is wide, you get a squeezed result; widen strip_width_pct or pad the overlay first.
• It's an additive frame transform, not a segmentation job - no models, no downloads. GPU-accelerated torch compositing.

Part of TrentNodes: ComfyUI Manager (search "Trent Nodes"), or clone + pip install -r requirements.txt. Small, deterministic, and it ports a genuine editor-workflow trick into the graph - there aren't many nodes doing this, and it's a nice one to have when you're delivering to an editor who likes their own keying.