ComfyUI Node
LLM Text
Process text using OpenAI, Gemini, Anthropic, or OpenRouter.
LLM Text
- text
◄system_promptYou are a helpful assistant. Return only the requested text.►
◄text►
◄provider▾►
◄api_key►
◄model►
◄max_tokens4096►
◄timeout120►
◄reasoning▾►
◄provider_slug►
◄routing_suffix▾►
◄cache_responsetrue►
CategoryLLM/Text
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| system_prompt | STRING | You are a helpful assistant. Return only the requested text. | — |
| text | STRING | Connect the external STRING to process. Convert to a widget for manual input. | |
| provider | COMBO | 4 options: OpenAI, Gemini, Anthropic, OpenRouter | |
| api_key | STRING | API key for the selected provider. Entered keys are stored in the browser and workflow. | |
| model | STRING | Enter a model ID or use Fetch Models and select a model. | |
| max_tokens | INT | 409616–1048576 | Output budget, including reasoning on some providers. |
| timeout | INT | 12010–1800 | — |
| reasoning | COMBO | Minimum selects the lowest supported reasoning setting. Explicit unsupported efforts return an error. | |
| provider_slug | STRING | OpenRouter only. Exact provider slugs, comma-separated. Restricts routing to these providers. | |
| routing_suffix | COMBO | OpenRouter only. Nitro prioritizes throughput; floor prioritizes price. Default preserves a suffix entered in Model. | |
| cache_response | BOOLEAN | true | Reuse the result for unchanged inputs. Turn off to request a new response every run. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| text | STRING | — |