▮ Gemini 3.1 Pro + Gemini 3.1 Flash Image · GMI Cloud

PICTURE COMPOSE

Place one image into another, end-to-end on GMI Cloud. The destination scene is preserved; only the inserted objects are added, with lighting, shadows, perspective and color grade matched so the result reads as a single in-camera photograph.

GMI key

01 · Destination (required)

Drop the scene to editPNG · JPEG · WEBP · click to browse

02 · Source objects (optional · up to 6)

Drop object imagesone object per image · click to browse

03 · Instruction (optional)

Resolution

Aspect

Compose = 1 vision call + 1 upload per image + 1 image job. Recompose reuses the analysis and uploads for an alternative take (1 image job).

Output

Composite appears hereUpload a destination, add objects or a note, then press Compose.

Model reasoning

Scene analysis

—

Object analysis (per source)

—

Composition prompt sent to the image model

—

Job info

—

VISION_LLM: google/gemini-3.1-pro-preview IMAGE_MODEL: gemini-3.1-flash-image-preview PROVIDER: gmicloud.ai

PICTURE COMPOSE

Build this on GMI Cloud