🔎 H3 Route Probe - LD

The diagnostic node that changes nothing

H3 Route Probe (🔎 H3 Route Probe - LD, under LD / PromptMaster / Video rail, in PromptMasterLD) is the strangest-feeling node in the pack, and one of the most useful when things go sideways: it shows the routing state that H3 Studio is about to dictate, and it changes nothing. No inputs, no settings, no wires in. You drop it on the canvas, queue once, and read what the router would do.

The reason it exists is the whole design bet of the pack's routing: the route nodes read H3 Studio's settings out of the submitted prompt at execution time instead of being hand-flipped. That kills the "two booleans disagree" class of bug, but it also means the state is invisible - nothing on the canvas shows you what mode Studio resolved to. When the wrong branch fires, your first question is "which branch is live?", and before this node the answer involved reading node source or hunting through the queue prompt. The probe is the answer to that question, on the canvas.

What you get

Three outputs, all read from the same submitted-prompt scan the route nodes use:

• is_multishot (BOOLEAN) - is the MULTISHOT tick in Studio on? This is the single most common thing to verify, because Studio returns positive=None when multishot is on, and routing a single-shot sampler from that state kills the queue with a None conditioning.
• is_ref (BOOLEAN) - is reference mode active?
• report (STRING) - the human-readable summary of the routing state it found, for a debug text node or just a look at the widget.

It's an output node, so it does nothing downstream and expects nothing upstream - think of it as the pack's version of a multimeter probe clipped onto the decision wire.

When you'd actually use it

Three cases. First: you just changed Studio's mode and you're not sure the router will agree - queue a probe first and confirm. Second: a render came out on the wrong rail and you want to know which setting caused it before you touch anything. Third: you're teaching yourself how the routing works - the probe makes the invisible decision concrete in a way that reading docs never does, because it reports exactly what the executing graph resolved to.

Install

cd ComfyUI/custom_nodes
git clone https://github.com/Lora-Daddy/PromptMasterLD
cd PromptMasterLD
pip install -r requirements.txt

Restart ComfyUI, or ComfyUI Manager → search "PromptMasterLD".

One honest note: because the probe reads the submitted prompt, it reports what Studio would do on the next run - it doesn't show you historical state or what a bypassed branch would have done if you'd wired it differently. If a render already happened, re-running the probe re-reads the current graph, which is what you want when you're iterating toward a fix rather than auditing the past.