These nodes cover video and audio generation, editing, and utility steps inside Workflows.

The Create Video node takes a prompt, an image, or both, and generates a video. The video's duration, aspect ratio, and content depend on the model chosen and the inputs provided.

To output multiple aspect ratios from a single workflow, combine video nodes with Composition Nodes earlier in the flow, using a separate video node per aspect ratio or resolution.

The Edit Video node uses the Kling O1 Video Edit model to modify specific areas of a video. As of this writing, Kling O1 is the only model available for this node.

To use it, connect a Create Video node or a reference video to the Edit Video node; Kling O1 can only edit existing videos. Include one or more images as reference groups, then reference them in your prompt as @Image1, @Image2, and so on, in natural language. For example: "Take the following video and replace the running character with @Image1. Then, replace the car they are chasing with @Image2. Keep all other elements of the video the same."

Videos sent to this node must be between 3 and 12 seconds long. For a 30 or 60 second stitched video, you will need multiple Edit Video nodes to cover the full length.

The Reframe Video node takes a source video (length limits vary by model; Ray 2 supports videos under 10 seconds) and reframes it into a different aspect ratio, using a prompt to define what fills the new space. This is useful for translating a square video into portrait or landscape. For the best results, though, we recommend generating a new video natively in each aspect ratio you need rather than reframing after the fact.

The Extract Frame node pulls the first or last frame from a video and passes it forward as a static image, usable the same way as any other image in a workflow. For example, you might generate an 8 second Veo 3.1 video in 9:16 and extract the final frame as a 1080x1920 portrait image.

To use it, connect a video generator or video reference into the node, then choose whether to extract the first or last frame. You must choose one; to extract both, use two separate Extract Frame nodes.

A common use is chaining multiple video generations together: generate an 8 second video, extract its final frame, and feed that into a second video generator's first frame input, so the second clip picks up exactly where the first left off. The two clips can then be joined with a Stitch Videos node.

The Trim Video and Trim Audio nodes isolate a smaller portion of a video or audio clip. Trimming is not guaranteed to be frame accurate: it searches from the beginning of the start second to the beginning of the end second, so a clip trimmed from second 0 to second 2 may not end on the same frame as one trimmed from second 2 onward. For frame accurate splitting, use the Video Timeline node, edit the video or audio directly, or use the Extract Frame node.

You can send multiple inputs into a single Trim node; all of them will be trimmed to the same start and end time.

The Audio Volume node adjusts an audio clip's volume, including fade in, fade out, and normalization. Use it on Reference Audio nodes or after extracting audio from a video. Like most workflow nodes, you can pass in multiple audio files to modify them all at once.