ComfyUI Node
Nano Banana Pro (Gemini Direct)
A ComfyUI node in 3DELAB with 14 inputs and 2 outputs.
Nano Banana Pro (Gemini Direct)
- images
- mask
- scribble
- images
- response_text
◄promptDescribe the image you want to generate...►
◄modelgemini-3-pro-image-preview►
◄operationGENERATE►
◄api_key►
◄service_account_json►
◄seed42►
◄aspect_ratio1:1►
◄resolution1K►
◄response_modalitiesIMAGE►
◄thinkingfalse►
◄system_promptYou are an expert image-generation engine. You must ALWAYS produce an image.►
Category3DELAB
Inputs (14)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Describe the image you want to generate... | The text description of the image you want to generate. |
| model | COMBO | gemini-3-pro-image-preview | Select the AI model to use (API Key for Gemini models to generate | JSON Key for Imagen variants for Inpaint, Outpaint, Background Swap). |
| operation | COMBO | GENERATE | Choose the operation mode: GEN (TXT2IMG|IMG2IMG|MULTIREFERENCE), INPAINT (Edit|Remove), OUTPAINT, BG_SWAP. |
| api_keyopt | STRING | Your Google AI Studio API Key. Required for Gemini models. | |
| service_account_jsonopt | STRING | Absolute path to your Vertex AI Service Account JSON key file. Required for Imagen 3 editing/inpainting. | |
| imagesopt | IMAGE | Input image for editing, inpainting, or image-to-image generation. | |
| maskopt | MASK | Mask image for inpainting (white = edit, black = keep). | |
| scribbleopt | IMAGE | Scribble or sketch image (transparent BG) for controlled editing. | |
| seedopt | INT | 420–18446744073709550000 | Seed for random number generation. |
| aspect_ratioopt | COMBO | 1:1 | The aspect ratio of the generated image. |
| resolutionopt | COMBO | 1K | Resolution of the output image (1K, 2K, 4K). |
| response_modalitiesopt | COMBO | IMAGE | Choose image only or image + reasoning text. |
| thinkingopt | BOOLEAN | false | Enable High thinking level (gemini-3.1 flash/pro only). |
| system_promptopt | STRING | You are an expert image-generation engine. You must ALWAYS produce an image. | System-level instructions to guide the model's behavior. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| images | IMAGE | — |
| response_text | STRING | — |