Nodes/ComfyUI-API-Toolkit/Gemini Inpaint
ComfyUI Node

Gemini Inpaint

A ComfyUI node in API Toolkit/Gemini/Image with 8 inputs and 1 output.

By IxMxAMAR·Created 5 months ago·Updated about a month ago· 1
Gemini Inpaint
  • image
  • mask
  • reference_image
  • image
api_key
modelgemini-3.1-flash-image-preview
custom_model
prompt
system_instruction
CategoryAPI Toolkit/Gemini/Image

Inputs (8)

NameTypeDefaultDescription
api_keySTRINGGemini API key. Leave blank to use GEMINI_API_KEY env var.
modelCOMBOgemini-3.1-flash-image-previewImage-capable Gemini model for inpainting.
custom_modelSTRINGOverride with a custom model ID.
imageIMAGEThe source image to inpaint.
maskMASKMask (white = fill this area, black = keep). Accepts any ComfyUI MASK — from SAM, Grounding DINO, alpha channels, etc.
promptSTRINGDescribe what should fill the masked area.
reference_imageoptIMAGEOptional reference image (e.g., object/face to place in the masked area).
system_instructionoptSTRINGSystem instruction for the inpainting model.

Outputs (1)

NameTypeDescription
imageIMAGE