ComfyUI Node
🎲 LLM Server Config
A ComfyUI node in prompt/wildcards with 6 inputs and 1 output.
🎲 LLM Server Config
- server
◄backendollama►
◄endpointhttp://localhost:11434►
◄modelllama3.1►
◄api_key►
◄temperature0.90►
◄llm_enabledtrue►
Categoryprompt/wildcards
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| backend | COMBO | ollama | 3 options: ollama, llamacpp, openai_compatible |
| endpoint | STRING | http://localhost:11434 | — |
| model | STRING | llama3.1 | — |
| api_key | STRING | — | |
| temperature | FLOAT | 0.900–2 | — |
| llm_enabled | BOOLEAN | true | Master switch for the LLM. When OFF, no node calls the LLM server: the Manager reuses its last generated template, and the Resolver fills __wildcards__ only from values already on disk (no value generation, no grammar pass) using the current template. Turn off to run the wildcard pipeline with the LLM server stopped. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| server | LLM_SERVER | — |