AS_GeminiCaptioning Node

User Guide

The AS_GeminiCaptioning node lets you generate a descriptive text prompt from an image using the Google Gemini API. Simply supply your image and adjust any optional text fields to tailor the output to your needs.

Inputs

• IMAGE (Required):
  The image you want to describe (e.g., JPEG, PNG). Connect your image input.

• PROMPT TYPE (Required):
  Choose between the preset styles "SD1.5 – SDXL" or "FLUX" to select the base style of the prompt.
  This determines which default reference text is used if you do not provide a custom reference.

• APY KEY PATH (Required):
  The file path to your API key needed for authenticating with the Google Gemini API.

• PROMPT LENGTH (Optional):
  An approximate word count for the final prompt.
  If left empty, no word count guidance will be added to the prompt.

• PROMPT REFERENCE (Optional):
  A sample text prompt format that serves as a reference for the generated description.
  If left empty, a default reference based on the selected PROMPT TYPE will be used.

• PROMPT STRUCTURE (Optional):
  A guideline for how the prompt should be organized (e.g., order of details such as building type, materials, location, etc.).
  If left empty, a standard structure will be applied automatically.

• IGNORE (Optional):
  Specific text or details that you want the node to exclude from the generated description.
  If left empty, nothing will be excluded from the prompt.

• EMPHASIS (Optional):
  Specific details or aspects you want to highlight in the generated description.
  If left empty, no additional emphasis instructions will be included.

• SAVE TO PATH (Optional):
  The directory where the generated text file should be saved.
  If left empty, the result will only be returned as output and not saved to a file.

• TXT NAME (Optional):
  The name for the saved text file. The node will automatically append a ".txt" extension if not provided.
  If left empty and a save path is provided, the file will be named "result.txt".

Outputs

• RESULT PROMPT:
  The text response generated by the Google Gemini API.

• REQUEST TEXT:
  The complete prompt text that was sent to the API.

• LOG:
  A log detailing the node’s execution steps, useful for troubleshooting and reference.

Required Libraries

• Pillow
• requests
• google-generativeai

pip install Pillow requests google-generativeai