Skip to content

Full Pipe

Full Pipe Concept

test

The idea of the full pipe nodes is to have a single connection between nodes and allow for easy inlining within full pipes through dedicated nodes and subgraphs. The above example shows a simple workflow for anima adding the CLIP NegPip node to the pipe via a subgraph and adding a lora with the Load LoRA (full-pipe) node.

These were made as an alternative to ComfyUI-Easy-Use pipe nodes, with the intention of being faster and storing metadata info in the pipe.

Loader (full-pipe)

Loads a selected checkpoint, sets CLIP skip, sets a seed, and optionally applies a LoRA stack from the input.

Split Loader (full-pipe)

Same as the loader node, but loads separate diffusion, CLIP, and VAE models.

Prompt (full-pipe)

This is a combined positive and negative prompt box that lets you use pre-defined prompt text. Features include:

  • Accepts LoRA syntax
  • Ignores # commented lines
  • Automatically applies quality tags, embeddings, style tags, character tags, and preset tags from Preset Manager.
  • Splits quality tags+embeddings, style tags, character tags, and the main prompt into separate conditionings and then concatenates the conditionings.

Simple Prompt (full-pipe)

This is a simpler version of the prompt node, made primarily for anima. It keeps the lora syntax and commenting, but omits the rest of the features. There’s a setting at the bottom of the node to convert the negative prompt to negpip, which requires the CLIP NegPip node from ComfyUI-ppm.

Base (full-pipe)

This is the base image generation node. By default it will use an empty latent with the dimensions defined by the node, but it also has an image input and denoise parameter for img2img generation.

Upscale (full-pipe)

This is an upscaling node that will upscale using lanczos and an upscale model (or none) to a target scale value and then resample the image.

Save (full-pipe)

This node saves the image with the ComfyUI workflow and A1111 metadata. I use a tool on my images that pulls the A1111 prompt (since pulling a prompt from a comfy workflow isn’t standardized in any way), so that’s the main focus of the node.