agnify nodes
agnify nodes list
Section titled “agnify nodes list”List the available pipeline node types, with one-line descriptions.
Serves the backend’s live registry (every node minus any disabled by a feature flag in this environment), so new nodes appear here automatically.
Usage
agnify nodes listagnify nodes models
Section titled “agnify nodes models”List the inference models this org may use, keyed by the node that selects them.
One key per model-selecting node type, or <node>.<backend> where a node
picks a model per backend. Each entry is the serving node’s own catalog,
relayed as-is: whatever it reports about a model — class labels, task, clip
window — comes through without this command naming it, so a new model, a new
field, or a new catalog needs no release here.
Models finetuned for other orgs are not listed — picking one would be refused when the run starts.
A catalog whose GPU service is dark comes back as an error string instead
of a list, so one dark service doesn’t hide the other. Every other failure
(an expired token, a bad org header) fails the command like any other
subcommand rather than exiting 0 with a menu the caller can’t trust.
Usage
agnify nodes modelsagnify nodes schema
Section titled “agnify nodes schema”Print the full ui_schema dict for one node type, including agent_hints.
Usage
agnify nodes schema NODE_TYPEArguments
| Name | Description |
|---|---|
NODE_TYPE | Node type, e.g. DFineInference |