I’m having AI assist me in creating a test workflow. For Advanced ControlNet, I installed v. 1.5.6. via manager Kosinkadink
However, it continues to reference a connector that doesn’t exist within Apply Advanced ControlNet” output called “CONDITIONING” . I did find that I have two Apply Advanced ControlNet’s … one of them is named exactly the same but has “(1”) at the end of the file name. where the node without (1) has positive and negative inputs, with positive and negative outputs. AI still recommends I avoid using (1) as it is “secondary”, but then continues to repeat having me connect KSampler to Apply Advanced ControlNet output to “CONDITIONING”. How am I supposed to do this if it doesn’t exist? I just installed ComfyUI on 8/23 and ran updates within the manager
Node 8: KSampler
-
Double-click the canvas > Type “KSampler” > Click to add (or right-click > “Add Node” > “sampling” > “KSampler”).
-
Why: This samples (generates) the image using the model and conditioned prompts with ControlNet.
-
What to expect: A box labeled “KSampler” appears with inputs like “model,” “positive,” “negative,” “latent_image” on the left, fields like “steps,” “cfg,” “sampler_name,” and output “LATENT” on the right.
-
Connect:
-
Drag from “MODEL” output of “Load Checkpoint” to “model” input of “KSampler”.
-
Drag from “CONDITIONING” output of “Apply Advanced ControlNet” to “positive” input of “KSampler” (for positive prompt with ControlNet applied).
-
Drag from “CONDITIONING” output of “CLIP Text Encode (Negative)” to “negative” input of “KSampler” (for negative prompt; if “Apply Advanced ControlNet” has a negative output, use that instead).
-
Would be more intuitive if they called it “Clip Text Encoded” 
but this is what we got and works ok after you get used. 
You can even use a ConditioningZeroOut node if your model don’t accept a negative prompt, like Flux. 