Batch Load Image Sequence

If you've got a folder of numbered frames - frame000000.png, frame000001.png, and so on, ripped from a video or rendered somewhere else - JWLoadImageSequence reads a run of them off disk and loads them as a single IMAGE batch you can push through a video workflow. This is the front door for frame-by-frame vid2vid: get your source video into frames, load them here, and now every frame is on one wire ready for AnimateDiff, ControlNet, or a straight img2img pass.

It's one of the reasons this pack ends up installed on so many machines - plenty of AnimateDiff and video-to-video workflows lean on it to pull frames in, so it rides along as a dependency of workflows people download. If you found it sitting in a graph you didn't build, this is almost certainly why.

How it works

You give it a filename pattern with a number placeholder, a starting number, and how many frames to grab. It walks the sequence - substituting the index into the pattern for each frame - and stacks the images into a batch. The default pattern ./frame{:06d}.png means "frame + a 6-digit zero-padded number + .png", so index 0 is frame000000.png, index 1 is frame000001.png. Match the pattern to how your frames are actually named.

The inputs and outputs that matter

• path_pattern (STRING) - the filename pattern with a {:06d}-style placeholder, default ./frame{:06d}.png. Adjust the digit count and folder to match your files.
• start_index (INT, default 0) - which frame number to begin at.
• frame_count (INT, default 16) - how many frames to load. 16 is a nod to typical AnimateDiff context lengths; raise it for longer clips.
• ignore_missing_images (enum: false / true) - if a frame in the range is missing, false errors out, true skips past the gap. Handy when a sequence has holes.
• Output: IMAGE - the frames as a batch.

How to install it

Part of jamesWalker55/comfyui-various, a pack of small utility nodes.

• ComfyUI Manager: search "Various ComfyUI Nodes by Type", install, restart. Red node in a downloaded workflow? Install Missing Custom Nodes.
• Manually:

  cd ComfyUI/custom_nodes
  git clone https://github.com/jamesWalker55/comfyui-various
  

  Restart afterward. No models, no dependencies.

Under the jamesWalker55 category after restart.

Common issues

• The pattern doesn't match my filenames. This is the number-one failure. If your frames are img_0001.jpg you can't use the default ./frame{:06d}.png - the prefix, the digit count ({:04d} vs {:06d}), and the extension all have to line up exactly. Rename your frames or edit the pattern to fit.
• It errors partway through the range. A missing frame with ignore_missing_images on false stops the load. Set it to true to skip gaps, or check that your start_index and frame_count stay within the frames that actually exist.
• Paths on a cloud/serverless host. The pattern points at the server's filesystem, and ./frame... resolves against ComfyUI's working directory, not your computer. On a hosted setup you need the frames to actually be on that machine (uploaded, or produced earlier in the pipeline) - a local path from your desktop won't resolve. Loading many high-res frames also eats VRAM/RAM fast, so keep batches sane.
• Node is red / missing. The pack isn't installed. Install via Manager or clone the repo, then restart.

Its partner is JWSaveImageSequence, which writes a processed batch back out as numbered frames - load, transform, save, then encode to video. That round trip is the backbone of a lot of frame-based pipelines.