WebUI Bridge Upscale Config

WebUI Bridge Upscale Config is the config tap for the pack's upscale module - same pattern as the ADetailer, ControlNet and SAM config nodes. The main Bridge node stores your upscale plan in its module_config JSON; this node extracts the upscale section so you can read it or pipe it elsewhere.

"Upscale plan" deserves the scare quotes because it's a genuine plan, not a single setting. The pack supports four different upscaling routes - hires-fix, latent upscale, Ultimate SD Upscale and tile - and each needs its own parameters: scale factor, upscaler method, second-pass steps and denoise for the hires routes, tile width/height and overlap for the tile routes. This node shows all of it in one readable JSON blob.

How it works

One input, two outputs. Feed it module_config and you get:

• config_json - the upscale section pretty-printed: mode (hires_fix/ultimate_sd_upscale/latent_upscale/tile), scale_by, upscaler, steps, denoise, tile_width, tile_height, overlap.
• enabled - whether the upscale module is switched on.

Like the other config nodes, enabled is the output that actually changes behavior downstream: the pack's Image Upscale and Latent Upscale nodes pass their input through untouched when it's false. If you think upscaling is running but your output is the same size, this is the first place to look.

Inputs and outputs

• module_config (required) - the settings JSON from the main Bridge node's module_config output.
• Outputs: config_json (STRING) and enabled (BOOLEAN).

How to install

Part of the ComfyUI-WebUI-Prompt-Bridge pack: ComfyUI-Manager → ComfyUI WebUI Prompt Bridge, or clone into custom_nodes and restart. No models for the node itself.

Common issues

Read this node's output before assuming a mode is what you think it is. ultimate_sd_upscale and tile modes are parameter storage only - the feature reference is explicit that the pack won't pretend a third-party tiled upscale flow is connected. Those modes need the Ultimate SD Upscale node pack (which the Bridge can install on demand) wired up yourself. The hires_fix and latent_upscale modes map to the pack's own Latent Upscale + second-sampler flow. And remember the classic upscaling category error this whole pack inherits: if your upscaler is set to a model name like "RealESRGAN," the pack's image/latent upscale nodes will fall back to an interpolation method rather than actually run the model - check config_json and you'll see exactly which route is real and which is a plan waiting for a model.