Gemini: Iterative image editing with Nano Banana

Gemini Try it →Watch full video →@KevinStratvert · Sep 2025

Demo summary

The video shows how to refine a generated image by changing the text on a sign and adjusting the lighting/weather conditions through follow-up prompts.

Step-by-step

Identify the specific element in the generated image you want to change
Type a refinement prompt in the text field to modify a specific detail like text on a sign
Enter a follow-up prompt to adjust environmental conditions such as lighting or weather
Review the updated image to ensure consistency with the original scene

Tips

Use iterative prompting to keep refining the image until it is exactly what you want
Provide specific names or text strings to swap out labels within the image

Highlights

“What makes this new model impressive is that you can keep editing and refining until it's exactly what you want.”

All demos from “Nano Banana AI Tutorial | How to Use Google’s Free Tool (Step-by-Step)”

1:250:34Generate images from text with Nano BananaThe user navigates to the 'Create Images' tool in Google Gemini and generates a storefront image from a simple text prompt using Nano Banana.Gemini· Text to Image
2:040:52Iterative image editing with Nano BananaCurrentThe video shows how to refine a generated image by changing the text on a sign and adjusting the lighting/weather conditions through follow-up prompts.Gemini· AI Inpainting
2:560:57Turn a photo into a 3D figurineKevin uploads a personal photo and uses a specific prompt to transform himself into a consistent 3D collectible action figure using Nano Banana.Gemini· AI Style Transfer
3:531:00Blend images and swap environmentsThe demo shows uploading two separate images (a person and a jacket) to blend them together and then changing the background environment to an office setting.Gemini· AI Image Combiner
Watch “Nano Banana AI Tutorial | How to Use Google’s Free Tool (Step-by-Step)” →

AI Inpainting

1:130:21Maintain character consistency with Gemini 2.5 FlashThe video demonstrates how Gemini 2.5 Flash (Nano Banana) can modify an existing image, such as changing a person's clothing to a suit, while keeping the original character's face consistent.Fireship
2:040:52Iterative image editing with Nano BananaCurrentThe video shows how to refine a generated image by changing the text on a sign and adjusting the lighting/weather conditions through follow-up prompts.@KevinStratvert
1:580:39Preview hairstyles with Gemini 2.5 FlashThe creator shows how to use Gemini to modify a person's hair in an uploaded photo, demonstrating a buzz cut and long blonde hair transformations.AI Samson
18:530:57Edit YouTube thumbnails with GeminiThe creator walks through upgrading a low-quality video screenshot into a thumbnail by prompting Gemini to change his expression, add studio lighting, and insert a phone into his hand.AI Samson
2:490:12Edit an existing AI image with natural languageThe user demonstrates how to modify a generated image by providing a follow-up prompt to make the sunset more dramatic using Nano Banana's editing capabilities.The AI Advantage
4:040:52Iterative editing on multiple AI graphicsThe user demonstrates how to modify a batch of generated images by adding a 'dabbing pose' and then removing specific floating objects like clovers using natural language commands.Philip Anders
10:520:43Precise object editing and lighting effectsGemini Omni is shown performing micro-edits such as changing a jacket's color and adding a light-tracking lens flare through text commands.AI Master
3:190:35Modify image colors using AI promptsThe demonstrator uses a text prompt to 'Change the color to purple' on a selected image, showing Google Mixboard's AI regenerating the asset in real-time.Learn AI Tools