E-commerce Visual Marketing
Generate product scene images, model try-on images, and marketing posters, reducing shooting and outsourcing costs. Turn a white-background product image into a complete set of channel assets with just a few API calls.
Typical Use Cases
- Scene composition: White-background image + prompt → home/outdoor/holiday and other scene images;
- Poster layout: Specify copy and style to directly generate marketing posters with text (Gemini image models have better text rendering results);
- Multi-aspect adaptation: Use the same asset with model suffixes to generate different formats—
-square(product card),-portrait(mobile),-landscape(banner).
Recommended Models
| Use Case | Recommendation |
|---|---|
| Scene images / try-on images (image-to-image) | gpt-image-2, gemini-3.1-flash-image |
| Text-based posters | gemini-3.0-pro-image, gpt-image-2-4k |
| Large-scale low-cost generation | gpt-image-2-fast series |
| Product short videos | omni_flash_landscape, veo_3_1_i2v_s (image-to-video) |
Implementation Notes
- See GPT-Image Image Generation and Nano Banana Image Generation for image-to-image calls;
- Images are billed per generation, and failures are not charged; before scaling up, calibrate prompts with small batches first;
- See Image and Video Generation Models for aspect ratio and resolution suffix rules。