Nodes/SSParallaxX/πŸ‘€ SS Image Uploader by SamSeen
ComfyUI Node

πŸ‘€ SS Image Uploader by SamSeen

A Load Image with a resize cap and an upload button

By SamSeenXΒ·Created about a year agoΒ·Updated 9 months agoΒ· 0
πŸ‘€ SS Image Uploader by SamSeen
    • IMAGE
    β—„imageβ–Ύβ–Ί
    β—„max_width512β–Ί
    β—„max_height0β–Ί

    This is the simplest node in SSParallaxX, and honestly it reads as "why does this exist?" at first glance. It's a Load Image with two extras: a max-resolution cap and a one-click upload button. Pick an image that's already in your ComfyUI/input folder (or hit Upload and drop one in), set max_width/max_height, and out comes a resized IMAGE tensor ready for the parallax node.

    Why reach for it instead of the built-in Load Image? Because the parallax generator is pure-Python slow and memory-hungry, so controlling resolution at the front of the pipeline matters more here than almost anywhere. Dropping a 2K render to 512 before it hits the depth model and the frame loop is the difference between a quick test and a coffee break.

    How it works

    No AI, no downloads, no depth model. It lists png/jpg/jpeg/webp/bmp files in your input directory, and a small bundled web extension adds the Upload button, which POSTs the file to ComfyUI's built-in /upload/image endpoint and refreshes the dropdown for you. Resizing preserves aspect ratio: max_width 512 keeps the height proportional, and max_height 0 means "auto" so only one dimension constrains. Set both to 0 and it passes the image through untouched. The resize is Lanczos, so you're not giving up quality on the way in.

    Inputs and output

    • image: dropdown of files in input/ - pick one, or use the Upload button to add it.
    • max_width (default 512, up to 4096) and max_height (default 0 = auto): your caps.
    • Output: a single IMAGE tensor [1,H,W,3], straight into Parallax_Generator_by_SamSeen's base_image (or anywhere else an IMAGE goes).

    Two small traps. If the input folder is empty you get a "none" entry and the node returns a blank 512Γ—512 tensor instead of erroring - a silent no-op if you forget to actually load something. And it's strictly one image at a time; no batches, so this is a single-input convenience node, not a bulk tool.

    Install

    Same pack, same steps as the rest of SSParallaxX - ComfyUI Manager β†’ search "SSParallaxX", or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/MrSamSeen/SSParallaxX.git
    cd SSParallaxX
    pip install -r requirements.txt
    

    Then restart and look under "πŸ‘€ SamSeen". The pack-wide caveat applies here too: the author migrated the repo to SamSeenX in late 2025 and main is currently stripped of code, so if a fresh clone comes back empty, pin an older commit or install via Manager. For this node specifically there's nothing heavy to download - no models, no ffmpeg - so it's the easiest piece of the pack to get running.

    CategoryπŸ‘€ SamSeen

    Inputs (3)

    NameTypeDefaultDescription
    imageCOMBO1 options: example.png
    max_widthINT51264–4096β€”
    max_heightINT00–4096β€”

    Outputs (1)

    NameTypeDescription
    IMAGEIMAGEβ€”