<aside> 🛠

Turn a Video Into an Image List — with Sampling and Frame-Count Controls

The opening step for any frame-by-frame video workflow: decode a video into an image list. Choose whether you want every frame or every Nth frame (for timelapses, previews, or lightweight processing), and cap the maximum number of frames to keep long videos from producing runaway lists. Pairs directly with Video Reassemble to close the loop, and with List Extract / List Inject to work on individual frames in the middle.

</aside>

What it does

Video Split reads an input video and emits an image list — one image per frame you want to keep. Two sliders control what actually makes it into the list:

• Sample every Nth frame — how much to decimate the video (1 = keep every frame; 2 = every other; 5 = every 5th; and so on)
• Maximum number of frames to output — a hard cap on the list size, so long videos don't produce unmanageable lists

Every image tool in the studio can then be applied to that list (batch-style or one-at-a-time via List Extract), and the finished list can be encoded back to video with Video Reassemble

Problem it solves

• Open the video pipeline loop – The entry point for any frame-by-frame video workflow
• Full-video enhancement with image models – Split → apply image tool per frame → Video Reassemble = enhanced video (upscaled, denoised, colorized, restored, stylized)
• Timelapse / speed-up effects – Sample every Nth frame to compress long footage into short, fast-motion clips
• Preview / thumbnail sheets – Decimate a video into a handful of representative frames for review or contact sheets
• Video-to-image dataset creation – Extract frames from a video for training, annotation, or ML datasets
• Cap resource usage on long videos – The max-frames slider prevents accidentally decoding a 20-minute video into 30,000 frames
• Keyframe-style sampling – Every-Nth sampling gives a coarse but even coverage of the video's content
• Bridge video → image tools – Any image-only tool in the studio (upscalers, editors, blurs, masking, generators) can consume the frames after a Video Split

Input/Output

<aside>

• Input: Single video

  • Format: Standard video formats (MP4, MOV, WebM, etc.)
  • Length / resolution: Any (long videos are safely capped by the max-frames slider)
  • Source: Any upstream video node or uploaded video file

  ebf8274b-d767-4c55-8bae-10e545d3fe16.mp4

</aside>

<aside>

• Output: Image list

  • Format: Standard image frames (PNG / JPEG) in list order
  • Ordering: Chronological — first item = first frame kept, last item = last frame kept
  • Length: min(⌈total_frames / N⌉, max_frames) — the smaller of "how many the sampling would yield" and "how many the cap allows"
  • Resolution: Matches the video's native frame resolution

  

</aside>

Configuration Options