- mbq_browser.py → mbq_viewer.py (git mv) - MetaViewApp → MBQViewerApp, MetaViewLogicMixin → MBQViewerLogicMixin - Window title and About dialog updated - README.md, comfy_nodes/README.md, CLAUDE.md: all "MBQ Browser" refs replaced with "MBQ Viewer"; stale increment=1 patch note removed - CLAUDE.md: drop stale "Future: wedge node" section; remove mbq_nodes.py from file map
ComfyUI-MBQWedge
A single ComfyUI node that sweeps a numeric parameter across a range, generating one image per value — analogous to a photographic wedge or exposure bracket.
Install (local / Phase 1)
Copy or symlink this comfy_nodes/ folder into your ComfyUI custom nodes directory:
ComfyUI/custom_nodes/ComfyUI-MBQWedge/
Restart ComfyUI. The node appears in the node list under MBQ → MBQ Wedge.
Usage
Add MBQ Wedge to your workflow and wire one of its two outputs:
| Output socket | Connect to | Example |
|---|---|---|
int_value |
Any INT input | KSampler → steps |
float_value |
Any FLOAT input | KSampler → cfg |
ComfyUI greys out incompatible socket types when dragging a link, so you can only connect the right type for the target parameter.
Inputs
| Input | Type | Description |
|---|---|---|
parameter_name |
STRING | Name of the swept parameter (auto-filled when you connect an output). |
start |
FLOAT | First value in the sweep |
stop |
FLOAT | Last value (inclusive) |
increment |
FLOAT | Step between values |
Node display widgets
- will produce → N iterations — live count of images that will be queued
- current → X.XX — the first iteration's value (cosmetic; updates as
startchanges)
Example: sweep steps 10 → 20 by 2
parameter_name = "steps"
start = 10
stop = 20
increment = 2
Wire int_value → KSampler steps. Click Queue → 6 separate jobs submitted automatically.
Example: sweep CFG 1.0 → 7.0 by 1.5
parameter_name = "cfg"
start = 1.0
stop = 7.0
increment = 1.5
Wire float_value → KSampler cfg. Click Queue → 5 separate jobs submitted automatically.
How Queue works with MBQWedge
When you click Queue, the MBQ JS extension intercepts the submission and expands it into
N individual jobs — one per sweep value. Each job has the swept value patched directly
into MBQWedge's inputs (start=V, stop=V, current=V), so the PNG produced
by each job carries the exact value in its embedded prompt chunk.
This is the same mechanism ComfyUI uses for per-image seed randomization.
Images appear in the output folder as each job completes — you don't need to wait for the full sweep before results start appearing.
MBQ Viewer integration
Open the generated images in MBQ. The viewer reads
MBQWedge.inputs.current from each image's embedded prompt chunk and overlays the swept
parameter and its value directly on the canvas (e.g. cfg: 3.5). The Wedge bulb in the
status bar lights green when a wedge is detected.
Because the value is embedded in the PNG itself, it remains correct if you move or rename the file.