Nodes/ComfyUI Video Segmentation Node/🐾MiaoshouAI Zip Compress
ComfyUI Node

🐾MiaoshouAI Zip Compress

Archive your video segments without leaving the graph

By miaoshouai·Created about a year ago·Updated about a year ago· 40
🐾MiaoshouAI Zip Compress
    • zip_filename
    filename_prefixComfyUI
    image_formatPNG
    password
    images_or_video_path

    When you've auto-cut a video into scenes and need to hand the results to someone - a client, a teammate, a second machine - the last thing you want is 47 loose MP4s to copy around. ZipCompress bundles whatever paths you give it into a single zip archive, all inside ComfyUI. It's the pack's little export utility, and it's the node you reach for at the end of a segmentation workflow.

    Inputs

    images_or_video_path is the input that matters, and its format is the one thing that will trip you up: it's a STRING containing one file path per line. That's not a format you'd guess from the name (it says "path", singular, and "images"), but it's exactly what "🐾MiaoshouAI Segment Video"'s path_string output produces - newline-joined paths. So the intended flow is Segment Video → path_string → ZipCompress, and you're archiving every segment in one shot. If you're feeding it a single path, just type it in.

    Also on the input side:

    • filename_prefix (default "ComfyUI") - sets the archive's name. The zip becomes {prefix}_{timestamp}.zip, written into the same folder as the first file you're archiving. Change the prefix if you want to tell your archives apart.
    • password (optional) - encrypts the archive. Honest warning: this uses Python's built-in zipfile encryption, which is the legacy ZipCrypto scheme, trivially breakable by modern standards. Fine for keeping a casual viewer out; don't put anything you actually need to protect through it.
    • image_format (PNG/JPG/WEBP/MP4/AVI/MOV) - looks important, does nothing. The source code literally comments it as "kept for interface compatibility." It does not convert anything, so set it to whatever and move on.

    Output

    A single zip_filename STRING - the absolute path to the finished archive. Wire it to a preview or Save node and you'll see the path right in the log. The archive uses DEFLATED compression at level 6: reasonable, but MP4 video is already compressed, so don't expect miracles. The point here is bundling, not squeezing.

    Two small behaviors: files that don't exist are silently skipped (with a console warning), so a stale path won't kill the whole archive. And if the input is empty, you get "" back rather than an error.

    Install

    Same as every node in this pack:

    cd ComfyUI/custom_nodes
    git clone https://github.com/miaoshouai/ComfyUI-Video-Segmentation
    cd ComfyUI-Video-Segmentation
    pip install -r requirements.txt
    

    …then restart ComfyUI - or ComfyUI Manager → Custom Nodes → search "MiaoshouAI" → Install. ZipCompress is a pure utility with no model and no extra dependencies, though you'll still absorb the pack's requirements.txt (which, notably, lists TensorFlow that the shipped code never imports). The real dependencies in this pack live in the Segment Video and Load TransNet Model nodes; this one just packages up what they produce.

    CategoryMiaoshouAI Video Segmentation

    Inputs (4)

    NameTypeDefaultDescription
    filename_prefixSTRINGComfyUI
    image_formatCOMBOPNG6 options: PNG, JPG, WEBP, MP4, AVI, MOV
    passwordSTRING
    images_or_video_pathoptSTRING

    Outputs (1)

    NameTypeDescription
    zip_filenameSTRING