The Extract Frame node is a utility node used to take the first or last frame from a video and then pass it as a static image to later nodes. For example, you may generate an 8 second video using Veo 3.1 in a 9:16 portrait aspect ratio and then export a 1080x1920 portrait image of the final frame of the video. Or, you may use the first frame of a video to generate variants later in the workflow. For the purposes of Layer's workflows, it will be used the same as any image.
To use the node, simply add it to your workflow graph and drag in a video node into it (either a video generator or a video reference). You may then select whether you would like to extract the first or last frame of the video. The output of the node will then be whichever frame you chose. You must extract the first or the last frame; If you would like to extract both, please use two nodes to pull the frames.
A common reason to use an Extract Frame node is to chain two or more video generations together. For example, you may generate an 8 second video using Veo 3.1, then extract the final frame from the output and feed it into a second video generator node as that node's first frame. This will allow you to start the second generator from the exact same position as the first generator's final frame, creating a continuous video that may then be placed together using the Stitch Videos node.
The Extract Frame node may be replicated in the inference form by dragging a video into the first frame portion of the UI. This will allow you to select a frame from the video to be used as the new video's first or last frame.