Seedance 2.5: FLF2V
Generate a smooth video transition between a user-provided start frame and end frame with Seedance 2.5, which automatically interpolates the motion, camera movement, and visual evolution while locking both endpoints. Ideal for shot transitions, storyboard previz, and before-and-after scene showcases like empty-to-packed or day-to-night shifts.
About
Seedance 2.5: FLF2V is a first–last frame to video workflow that generates a smooth, locked-endpoint transition between two user-provided images. It uses two LoadImage nodes to bring in your start and end frames, then passes them to ByteDance2FirstLastFrameNode, which calls the Seedance 2.5 model to synthesize the in-between motion and appearance changes while preserving both endpoints. Finally, SaveVideo encodes the returned frame sequence into a playable video file.
What you provide
- Reading the workflow
Runs in ComfyUI. This workflow can't run on this page. Open it in Comfy Cloud with everything loaded, or download it for your own ComfyUI.
Try in CloudQuestions & answers
Yes—use matching resolution and aspect ratio for stable interpolation. Crop or pad one image to match the other before sending both into ByteDance2FirstLastFrameNode.
Adjust the total frames generated (if exposed by ByteDance2FirstLastFrameNode) and/or the FPS in SaveVideo. More frames or lower FPS lengthens the clip; fewer frames or higher FPS shortens it.
Use clean, minimally compressed inputs with consistent framing. Align subjects similarly across both images, avoid drastic composition jumps, and, if available, tweak the node’s interpolation/quality settings or try a different seed.
No. SaveVideo writes a silent video. Add music or sound effects later in a video editor.