ComfyUI Node
OpenAI Vision
A ComfyUI node in ERPK/OpenAI with 10 inputs and 1 output.
OpenAI Vision
- image
- client
- analysis
◄promptDescribe this image in detail.►
◄seed-1►
◄modelgpt-5.5►
◄detailauto►
◄max_tokens4096►
◄temperature0.40►
◄reasoning_effortnone►
◄verbositydefault►
CategoryERPK/OpenAI
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | Image(s) to analyze (ComfyUI tensor) | |
| prompt | STRING | Describe this image in detail. | Question or instruction about the image(s) |
| seed | INT | -1-1–2147483647 | Seed for reproducible outputs (best-effort). Randomizes by default. |
| clientopt | OPENAI_API_CLIENT | OpenAI API client from OpenAI API Config node (optional if API key is configured in Settings) | |
| modelopt | COMBO | gpt-5.5 | OpenAI model to use for vision analysis |
| detailopt | COMBO | auto | Image detail level (low=faster/cheaper, high=more detailed) |
| max_tokensopt | INT | 4096256–16384 | Maximum length of analysis |
| temperatureopt | FLOAT | 0.400–2 | Creativity level (lower=more factual) |
| reasoning_effortopt | COMBO | none | Reasoning depth for o-series and gpt-5.x reasoning models. Ignored by non-reasoning models. |
| verbosityopt | COMBO | default | Output verbosity for gpt-5.x models. 'low' produces terse responses, 'high' produces more detailed ones. Distinct from max_tokens — shapes style, not the hard length cap. 'default' lets the model choose. Silently ignored by older models that do not accept verbosity. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| analysis | STRING | — |