RH Node Info
RH Node Info is how you inject local values into a cloud workflow
- previous_list
- node_info_list
Here's the situation RH Node Info solves: you've built a workflow on RunningHub's website - models loaded, samplers wired, saved. It works. But now you want to run that exact workflow from your local ComfyUI with your prompt, your seed, this time. The workflow ID is fixed. The parameters have to come from somewhere. That's this node: it's the only way the pack lets you override what a specific node inside the cloud workflow does.
Each RH Node Info instance is one parameter override. You tell it three things - which node in the cloud workflow, which field on that node, and what value - and it emits a single entry in RunningHub's nodeInfoList format:
{"nodeId": "6", "fieldName": "text", "fieldValue": "a cat in a spaceship"}
The three fields that matter
node_id- thenodeIdof the target node in your RunningHub workflow. This is the fiddly one: you have to know the ID, and RunningHub doesn't exactly advertise it. The API docs' nodeInfoList guide is the official way to figure out which nodeId maps to which workflow node. Get it wrong and the override silently targets the wrong node.field_name- the parameter name on that node, e.g.text,seed,image. The tooltip lists those three because they're the ones you'll use 95% of the time.field_value- the value. It's a multiline box, so big prompts are fine.
Chaining
The sneaky bit is previous_list (optional). Each node emits an ARRAY, and if you connect one node's output into another's previous_list, the second appends its entry onto the first - so the chain accumulates a list of overrides. One RH Node Info per parameter you want to control, daisy-chained, all feeding one Execute node. This is exactly how you set a prompt and a seed and an image in a single cloud run. (Compare with the Model nodes like RH Text to Image, which hardcode the node IDs for you - RH Node Info is the manual, any-workflow version of the same idea.)
Wiring it up
The chain output, node_info_list, plugs into the node_info_list input on RH Execute Workflow or RH Execute AI App. Yes, the input and output share a name - that's how the chain connects to itself and how it lands in the Execute node.
A couple of practical notes from the community trenches: RunningHub workflows don't always expose every node's fields for override, and if a nodeId in your list doesn't exist in the workflow, the platform can throw the whole submission. So start with one override, confirm it lands, then build the chain. And remember this node is pure plumbing - no API key, no network call; it just builds a data structure. If your Execute node comes back with output that ignores your prompt, check the nodeId, because that's the failure that doesn't say "failed" anywhere.
Install
Same pack, same recipe:
cd ComfyUI/custom_nodes/
git clone https://github.com/liangzheng1128/ComfyUI-RunningHub
cd ComfyUI-RunningHub
pip install -r requirements.txt
or ComfyUI Manager → search "RunningHub". Restart and it's under the RunningHub category alongside the rest.
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| node_id | INT | 00–999999 | The nodeId of the target workflow node |
| field_name | STRING | The parameter field name (e.g. 'text', 'seed', 'image') | |
| field_value | STRING | The parameter value to set | |
| previous_listopt | ARRAY | Connect another RH_NodeInfo output here to chain entries |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| node_info_list | ARRAY | — |