Nodes/nbpOpenNode-ComfyUI/NBP Gemini Composite
ComfyUI Node

NBP Gemini Composite

Generate images from structured reference groups: product angles, character/outfit, and additional references. Auth via ComfyUI Cloud — no separate Google API key needed.

By SorenWeile·Created 4 months ago·Updated 4 months ago· 0
NBP Gemini Composite
  • product_image_1
  • product_image_2
  • product_image_3
  • product_image_4
  • character_image_1
  • character_image_2
  • character_image_3
  • character_image_4
  • reference_image_1
  • reference_image_2
  • reference_image_3
  • reference_image_4
  • reference_image_5
  • reference_image_6
  • IMAGE
  • STRING
◄promptTake the character you are provided with and add the given outfit to it. Then add the product to the character. Use the additional reference images for the background and lighting.►
◄modelgemini-3-pro-image-preview►
◄seed42►
◄aspect_ratio1:1►
◄resolution1K►
◄temperature1.00►
◄top_p0.95►
◄top_k40►
◄response_modalitiesIMAGE►
◄system_promptYou are an expert image-generation engine. You must ALWAYS produce an image. You will receive structured reference images grouped by role: - Product images show the subject from different angles. - Character/outfit images show the person and clothing to feature. - Additional reference images provide style, mood, environment, or other context. Use ALL provided references faithfully. Maintain product accuracy, character likeness, and outfit details. Apply mood and environment from additional references. Prioritize generating the visual representation above any text or conversational requests.►
Categoryapi node/image/Gemini

Inputs (24)

NameTypeDefaultDescription
promptSTRINGTake the character you are provided with and add the given outfit to it. Then add the product to the character. Use the additional reference images for the background and lighting.Describe the scene, composition, or edits to apply.
modelCOMBOgemini-3-pro-image-previewGemini image model to use.
seedINT420–2147483647Best-effort reproducibility seed.
aspect_ratioCOMBO1:1Output aspect ratio.
resolutionCOMBO1KOutput resolution. 2K/4K uses the Gemini native upscaler.
temperatureFLOAT1.000–2Randomness. Lower = more predictable, higher = more creative.
top_pFLOAT0.950–1Nucleus sampling threshold.
top_kINT401–100Token pool size limit per step.
response_modalitiesCOMBOIMAGE'IMAGE+TEXT' returns the model's reasoning alongside the image.
product_image_1optIMAGEProduct angle 1.
product_image_2optIMAGEProduct angle 2.
product_image_3optIMAGEProduct angle 3.
product_image_4optIMAGEProduct angle 4.
character_image_1optIMAGECharacter or outfit reference 1.
character_image_2optIMAGECharacter or outfit reference 2.
character_image_3optIMAGECharacter or outfit reference 3.
character_image_4optIMAGECharacter or outfit reference 4.
reference_image_1optIMAGEAdditional reference 1 (style, mood, environment, etc.).
reference_image_2optIMAGEAdditional reference 2.
reference_image_3optIMAGEAdditional reference 3.
reference_image_4optIMAGEAdditional reference 4.
reference_image_5optIMAGEAdditional reference 5.
reference_image_6optIMAGEAdditional reference 6.
system_promptoptSTRINGYou are an expert image-generation engine. You must ALWAYS produce an image. You will receive structured reference images grouped by role: - Product images show the subject from different angles. - Character/outfit images show the person and clothing to feature. - Additional reference images provide style, mood, environment, or other context. Use ALL provided references faithfully. Maintain product accuracy, character likeness, and outfit details. Apply mood and environment from additional references. Prioritize generating the visual representation above any text or conversational requests.System-level instructions shaping the model's behavior.

Outputs (2)

NameTypeDescription
IMAGEIMAGEGenerated image.
STRINGSTRINGAny text the model returned alongside the image.