MNeMoNiCuZ-ComfyUI-mnemic-nodes/web/docs/MNeMiC_ResolutionSelector.md

📐 Resolution Image Size Selector

Picks an output resolution from a preset, your own saved preset, custom numbers, or an incoming image — then optionally multiplies, clamps, snaps and swaps it. Hands back the numbers plus a matching empty latent.

Inputs

• preset — Model-specific resolution presets, grouped by model. Custom uses the width/height widgets.
• preset_user — Your own presets. None means "ignore this".
• custom_width / custom_height — Used when preset is Custom and no higher-priority source applies.
• image (optional) — Connect an image to take its dimensions.

Advanced:

• multiply — Scales the final result. Negative values flip width and height.
• swap_width_and_height — Swaps the two.
• image_min_length / image_max_length — When an image is connected, clamp its shortest / longest side. 0 disables the clamp.
• snap_to_nearest + snap_resolution — Round both dimensions to a multiple of snap_resolution. 0 disables snapping.
• batch_size — How many latents to make in the batch.

Outputs

• width / height — The final numbers, in pixels.
• latent — An empty latent of that size, batch_size deep.

Priority

Highest wins:

1. image (optional) — if connected
2. preset_user — if not None
3. preset — if not Custom
4. custom_width / custom_height

The multiply, clamp, snap and swap steps then run on whatever that produced. Dimensions are never allowed below 64 pixels.

Your own presets

User presets live in nodes/resolution_selector/user_resolution.json and are created with a Favorites group the first time the node loads. Edit that file to add your own; entries are grouped by the top-level key and shown as Group: name. The model presets are in preset_resolution.json next to it.

{
    "Favorites": [
        {"name": "[Square] 768x768 1:1", "width": 768, "height": 768},
        {"name": "[Portrait] 832x1216 13:19", "width": 832, "height": 1216}
    ]
}

Reload ComfyUI after editing — the lists are read when the node is registered.