Video nodes

Video nodes generate and process moving pictures. They can generate straight from text or use images and videos as references, and their toolbar has video post-processing tools.

Create a video node

• Choose “Video” in the “Add node” menu. By default it generates video from text;
• You can switch modes in the prompt bar: text to video, Omni reference, image to video, first and last frames, and image reference (some models also have doc to video). If you connect an image or video while the node is still in text to video, it switches to “Omni reference” automatically, and an image-to-video node also switches to “Omni reference” when you connect a second image;
• Choose “Video” from the “+” beside an image node to create a video node that generates as image to video, connected automatically; switch to Omni reference or other modes in the prompt bar. Choose “First & last frame” to create a first-and-last-frame node with this image as the first frame, then connect another image as the last frame. “Avatar” in the same menu creates an avatar node.
• Extend video, edit video, segment remake and lip sync under “Video processing” are created from the “Add node” menu or the “+” beside a video node. For a draft made with “Draft preview” turned on, you can click “Make final video” on the draft within 7 days to generate the final version with the same model at the final resolution you choose; confirm the price, then submit.

Video editing tools

With a video node selected, these are the main tools on its toolbar (with whether each is billed):

• Video upscale: choose a target resolution on the toolbar (720p / 1080p / 2K / 4K, 1080p by default) and click “Generate”. A new node is created downstream, where you pick a model, check the price and submit. The price and pricing unit depend on the model you choose;
• Trim video: drag the in and out points to keep one section (free);
• Shot breakdown (keyframes, motion and music references): creates a shot breakdown node that splits the video at its cuts, writes a shot-by-shot table, and places three groups on the canvas: a keyframe and a clip for each shot, plus the original audio track (the model analysis is paid; frames, clips and audio are free);
• Smart Cut (multi-track): opens the multi-track timeline editor (exporting is free);
• Split audio & video: separates the clip into a silent video and an audio track in one click (free);
• Mix audio: choose another audio clip to mix in or replace the sound (free);
• Capture frame: grabs the current frame, the first frame, the last frame, or both the first and last frames (free);
• Vocal separation: keep only the vocals, or only the background sound (free);
• Remove subtitles: choose the subtitle type (dialogue subtitles or full-screen subtitles) and processing quality (Refined or Smart), then remove the subtitles from the video (paid);
• Smart highlights and Video summary: appear on the toolbar when a model is available for them (paid).

After you click “Generate” for video upscale or subtitle removal, a node with those settings filled in is created downstream; pick a model on that node, confirm the price and submit.