Generate fruit characters with Higgsfield Nano Banana 2

Higgsfield Try it →Watch full video →Money4Life · Apr 2026

Demo summary

The creator demonstrates how to use the Nano Banana 2 model within Higgsfield to generate stylized fruit characters by entering descriptive prompts and adjusting aspect ratios.

Step-by-step

Navigate to the image section in Higgsfield
Select the Nano Banana 2 model
Enter a descriptive prompt detailing the character's age, outfit, lighting, and personality
Set the aspect ratio to 9:16 for short-form or 16:9 for YouTube
Increase the quality setting for better details
Click generate
Select the best variation based on lighting, angle, and expression

Options

9:16 aspect ratio
16:9 aspect ratio

Tips

Describe your character as if they are part of a reality show for better results
Choose the variation with the most natural lighting and expression to ensure a high-quality final video
Repeat the generation process to create a full cast of characters

All demos from “How I Made Viral Fruit Love Island AI Videos with Higgsfield (Full Tutorial)”

1:090:53Generate fruit characters with Higgsfield Nano Banana 2CurrentThe creator demonstrates how to use the Nano Banana 2 model within Higgsfield to generate stylized fruit characters by entering descriptive prompts and adjusting aspect ratios.Higgsfield· Text to Image
2:091:54Build scenes in Higgsfield Cinema StudioThe user shows how to upload generated characters as references into Cinema Studio, select environments, and describe character interactions to create a static base scene.Higgsfield· AI Image Generator
4:031:42Animate AI scenes with Kling 3.0The narrator demonstrates converting a static scene into a video by selecting the Kling 3.0 model, adding motion prompts, and setting the duration and quality to 1080p.Higgsfield· Image to Video
Watch “How I Made Viral Fruit Love Island AI Videos with Higgsfield (Full Tutorial)” →

Text to Image

6:400:38Generate a cinematic location in HiggsfieldThe creator uses Cinema Studio 2.5 to generate a high-quality image of a rooftop over a war-torn city by inputting specific emotional and lighting prompts.Youri van Hofwegen
2:280:25Generate cinematic locations with Higgsfield AIThe demo shows how to generate a 'hotel corridor' location with specific cinematic lighting and depth using a text prompt.Higgsfield AI
9:510:54Generate a cinematic location in HiggsfieldThe user demonstrates generating a specific environment by prompting for an 'old western town' and selecting from multiple AI-generated variations to save as a location element.David Manning
6:000:41Generate AI images with Higgsfield Nano Banana ProThe user navigates to the Nano Banana Pro model in Higgsfield, pastes the prompt from ChatGPT, and generates a high-quality base image of strawberry characters.Taro Creates
3:550:39Generate fruit images in HiggsfieldThe narrator demonstrates selecting the Nano Banana 2 model in Higgsfield, pasting a prompt from ChatGPT, and configuring settings like 9:16 aspect ratio and 2K quality to generate strawberry character images.AI Century
52:131:21Generate images with GPT Image 2The creator demonstrates the core image generator by creating a 1970s New York fashion model photo using the GPT Image 2 model.Artturi Jalli
3:510:32Generate images via Telegram with Higgsfield AIThe user sends a text prompt to the Higgsfield AI agent within the Telegram app to generate a futuristic image using the GPT 2.0 model.metricsmule
1:090:53Generate fruit characters with Higgsfield Nano Banana 2CurrentThe creator demonstrates how to use the Nano Banana 2 model within Higgsfield to generate stylized fruit characters by entering descriptive prompts and adjusting aspect ratios.Money4Life
1:190:51Generate a start frame image in Higgsfield AIThe user demonstrates how to generate an initial image in Higgsfield AI by selecting a camera (Red V Raptor), lens (Helios), and focal length before entering a text prompt.InstantHowTo
3:151:35Generate AI images with Higgsfield CanvasThe user inputs a detailed cinematic prompt into a node and uses the 'GPT image 2' model to generate a high-quality image directly on the canvas.Mr.Endatah