Nodes/ComfyUI-ConDelta/Multi-Dimensional Prompt Travel
ComfyUI Node

Multi-Dimensional Prompt Travel

Several travel axes in one prompt field

By envy-ai·Created 2 years ago·Updated 11 months ago· 209
Multi-Dimensional Prompt Travel
  • clip
  • CONDITIONING
prompt

The plain PromptTravel node in this pack blends exactly two prompts along one axis - a single travel amount moving you from prompt_a to prompt_b. MultiDimensionalPromptTravel is built for a more ambitious version of the same idea: multiple travel axes packed into one prompt field, so you could in principle move along several independent dimensions at once (say, style along one axis and subject along another) instead of chaining several two-prompt travel nodes together.

Where the docs run out - read this before you build a workflow around it

Here's the honest gap: the node's own tooltip on prompt says "Use special formatting for multi-dimensional travel," and that's the extent of what's documented anywhere the pack ships. Neither the README nor the schema spells out the actual syntax - what delimiters separate dimensions, how you specify a travel amount per axis, or how many axes are supported. If you're trying to use this node for the first time, your best move is to open the pack's sample_workflows folder on GitHub (github.com/envy-ai/ComfyUI-ConDelta) and find a workflow that uses it, then reverse-engineer the formatting from a working example rather than guessing at syntax. This is a genuine documentation gap in the pack itself, not something specific to this write-up.

The inputs and outputs that matter

  • prompt - multiline text, where the special multi-dimensional formatting (whatever it turns out to be) lives.
  • clip - the CLIP model used to encode.

Output is a single CONDITIONING, same as every other node in this pack.

How to install it

Via ComfyUI Manager: search "ComfyUI-ConDelta", install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/envy-ai/ComfyUI-ConDelta

then restart. No model downloads for this node.

Common issues & troubleshooting

If the node just encodes your prompt normally with no visible travel effect, you probably haven't hit the special formatting syntax. Without it, this likely behaves like a plain CLIP Text Encode - a valid but unremarkable result. Check the sample workflows before concluding the node is broken.

Start with the plain PromptTravel node if you only need one axis. It's simpler, better understood from its schema alone, and doesn't ask you to reverse-engineer any formatting. Reach for the multi-dimensional version once you've confirmed you actually need more than one travel direction at a time and you're willing to dig into the sample workflows to learn the syntax.

This is exactly the kind of thing worth testing on a small, cheap generation first. Since the formatting isn't documented, experiment with a short prompt and a fast model before committing it to a long or expensive workflow.

Categoryconditioning

Inputs (2)

NameTypeDefaultDescription
promptSTRINGThe text to be encoded. Use special formatting for multi-dimensional travel.
clipCLIPThe CLIP model used for encoding the text.

Outputs (1)

NameTypeDescription
CONDITIONINGCONDITIONING