AIHub Action Set Project Config String

AIHub Action Set Project Config String stores a piece of text in the client project's config - a model id, a prompt fragment, a chosen style label, an arbitrary note. It's the string member of the pack's four config setters (boolean, float, integer, and this one), and like the others it does exactly one thing: write a value under a dotted field path so that a later workflow or the client app can read it back. No output, no file, just state.

It belongs to otavanopisto's ComfyUI-aihub-workflow-exposer, the pack that lets external applications drive ComfyUI as their generation engine. The "project" is client-owned state - a folder with a config.json in the basic case - and string config values are how a pipeline remembers things that aren't numbers: which checkpoint was used, which LoRA set, what the client should label the next output.

How it works

On execution it sends a single message over the pack's websocket (port 8111): {"type": "SET_CONFIG_VALUE", "field": "<field>", "value": "<value>"}. Dots in the field path create nested levels in the config. The client handles persistence however its project store wants to - that's deliberately out of the server's hands.

The inputs that matter

• field (required) - the dotted path, e.g. model.last_used or project.tags.
• value (required) - the string to store. Defaults to empty if it's ever null.

No outputs. Pure side effect.

Installing it

ComfyUI Manager (search "ComfyUI-aihub-workflow-exposer"), or:

cd ComfyUI/custom_nodes
git clone https://github.com/otavanopisto/ComfyUI-aihub-workflow-exposer
# restart ComfyUI

No requirements, no downloads.

Where people get burned

The same path-contract rule as every config setter: the dotted path you write must be the exact one the reader uses, or the value is invisible. Strings are also the easiest to trip over with accidental whitespace or newlines baked into the value - what you type into a multiline widget is exactly what gets stored. And once more, with feeling: without a project context and a connected aihub client, this node is a silent no-op.