Nodes/ComfyUI-Addoor/RunningHub Node Info
ComfyUI Node

RunningHub Node Info

Set inputs on a cloud workflow

By Eagle-CN·Created 2 years ago·Updated about a year ago· 75
RunningHub Node Info
  • previous_node_info
  • NODEINFOLIST
node_id
field_name
field_value
comment

This one only makes sense once you know what the whole "RHAPI" corner of ComfyUI-Addoor is for. RunningHub is a cloud ComfyUI service - you host a workflow on their servers and fire it off through an API instead of running it on your own GPU. The Addoor RHAPI nodes turn your local graph into a client for that API, and RunningHubNodeInfo is the node that says "when you run my cloud workflow, set this field on that node to this value."

Think of it as the override slip. A cloud workflow has a fixed shape, but you want to change the prompt, the seed, or a filename each run. This node builds that instruction.

How it works

Every node in a ComfyUI workflow has a numeric id and named widgets. RunningHubNodeInfo takes four things: node_id (which node in the remote workflow), field_name (which widget on it), field_value (what to set it to - this one is multiline, so long prompts are fine), and comment (a label just for you, so you remember what this override does). It packs those into a NODEINFOLIST and hands it out.

The clever part is previous_node_info, the one optional input. It's also a NODEINFOLIST. You chain these nodes: the first one's output feeds the second one's previous_node_info, the second feeds the third, and so on. Each node adds its override to the growing pile. At the end you've got one list holding every field you want to change, ready to attach to whatever node actually launches the remote run.

So the pattern is a little daisy chain - one link per field you're overriding.

The inputs that matter

  • node_id - must match the id of the node in the remote workflow, not your local graph. You get these from the workflow's API-format JSON (in ComfyUI, Save → export as API), or from RunningHub's interface.
  • field_name - the exact widget name, e.g. text, seed, ckpt_name. Spelling counts.
  • field_value - the value. Multiline, so paste whole prompts here.

The single output is NODEINFOLIST, which either chains into the next NodeInfo or feeds the run.

Where people get stuck

The whole thing is only as good as your node_id and field_name. There's no autocomplete pulling the remote workflow's schema for you - if you guess the id or misspell the widget, the override silently does nothing and you spend twenty minutes wondering why your prompt didn't change. Export the target workflow as API JSON and read the ids straight off it before you wire anything. That's not optional; it's the actual work.

It's also worth remembering this is a remote-control setup. Nothing renders on your machine. If you just want to generate images locally, none of the RHAPI nodes are what you want - they're for scripting someone else's cloud.

Installing ComfyUI-Addoor

Easiest path is ComfyUI Manager: open Install Custom Nodes, search "ComfyUI-Addoor" (or just "Addoor"), install, restart. Manual works too:

cd ComfyUI/custom_nodes
git clone https://github.com/Eagle-CN/ComfyUI-Addoor.git
cd ComfyUI-Addoor
pip install -r requirements.txt

Restart ComfyUI and the nodes show up under the 🌻 Addoor menu (RHAPI submenu for this one). The pack's docs are Chinese-first, so the READMEs won't hold your hand much - but the RHAPI nodes are simple enough that the field names tell the story.

Category🌻 Addoor/RHAPI

Inputs (5)

NameTypeDefaultDescription
node_idSTRING
field_nameSTRING
field_valueSTRING
commentSTRING
previous_node_infooptNODEINFOLIST

Outputs (1)

NameTypeDescription
NODEINFOLISTNODEINFOLIST