ComfyUI Node
⭐ Star Ollama Prompt Helper
Use a local Ollama model to create or refine prompts for image generation.
⭐ Star Ollama Prompt Helper
- image
- result
- think_output
◄local_addresshttp://127.0.0.1:11434►
◄model▾►
◄keep_alive5m►
◄system_prompt_presetCustom►
◄system_prompt►
◄prompt►
◄allow_thinkingfalse►
◄temperature0.80►
◄seed0►
Category⭐StarNodes/Prompts
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| local_address | STRING | http://127.0.0.1:11434 | Ollama server address. |
| model | COMBO | Select a model. Click refresh to load available models from the server. | |
| keep_alive | COMBO | 5m | How long the model stays loaded in memory after generation. '-1' = forever, '0' = unload immediately (slow on repeat runs). |
| system_prompt_preset | COMBO | Custom | Choose a preset system prompt or 'Custom' to write your own. Presets are loaded from the shared text storage (json/startext.json). |
| system_prompt | STRING | Custom system prompt (used when preset is 'Custom'). | |
| prompt | STRING | Your prompt text to send to the model. | |
| allow_thinking | BOOLEAN | false | If True, outputs the reasoning process. If False, instructs the model to skip reasoning to speed up generation. |
| temperature | FLOAT | 0.800–2 | Higher values produce more creative output. |
| seed | INT | 00–2147483647 | — |
| imageopt | IMAGE | Optional image input for vision language models. Refer to it in your prompt as 'this image'. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| result | STRING | — |
| think_output | STRING | — |