Nano Banana Pro stands out as the top image generation model right now. It creates high-res images with the same characters every time. Text looks perfect, physics make sense, and you can even hit 4K quality.

A woman in a dress and hat holds a coffee cup in a shop. The result? A pro photo with readable signs. Now swap that cup for ice cream and the shop for a beach. Same face, necklace, hair—everything matches.

N8N lets you automate workflows like a boss. This guide walks you through adding Nano Banana Pro to N8N. You’ll start simple, then gain full control for edits and multi-image combos. Get ready for consistent characters and product promos that wow.

Section 1: Accessing the Source – Setting Up API Credentials

Getting Your API Key from Google AI Studio

Go straight to Google for Nano Banana Pro. Skip third-party sites to dodge extra costs. Head to aistudio.google.com.

Click “Get API key.” Hit “Create API key.” Name it something like “Nano Banana Pro Tutorial.”

Pick a Google Cloud project. No projects? Create one fast. Give it a name, select it, then make the key. Google often tosses in $100 free credits on signup. Test the model right there before moving on.

Establishing Connectivity: N8N Setup Options

You need an N8N instance to build workflows. Self-host on your PC with my linked guide. Or try N8N Cloud at n8n.io—starts at 24 euros a month with extras.

For cheap self-hosting, use Hostinger VPS. Grab KVM2, pick your spot, and under OS, choose Applications > N8N. It installs in one click. Use code “Leon” for 10% off, especially on Black Friday deals at $5/month.

In N8N, go to Settings > License Key. Paste the free key from email. Activate it. Now you unlock pro features. Create a new workflow named “Nano Banana Pro Tutorial.” Add a trigger node. You’re set.

Section 2: The Simplest Integration – Initial Workflow Setup

Utilizing the Native Google Gemini Node (Limited Control)

Start easy with N8N’s Google Gemini node. Search “Google Gemini” and pick “Generate an Image.”

Create credentials. Name it “Nano Banana Pro Tutorial.” Paste your API key from AI Studio. Save. Green check means success.

Switch model to “by ID.” Enter models/gemini-3-pro-image-preview. Add a prompt like “a dog eating a hot dog.” Execute. Boom—image in seconds. Basic works, but you want more.

The Limitation of Simplicity: Why Custom API Calls are Necessary

This node skips key tweaks. No aspect ratio picks. No res choices. Can’t edit images or mix multiples.

Time for HTTP Request node. It opens full power. Control 4K output, edits, everything. Ditch limits.

Section 3: Advanced Integration – Unlocking Full Control with HTTP Requests

Configuring the HTTP POST Request for Image Generation

Add a form node named “Create Image.” Make a textarea for “prompt.” Required.

Connect HTTP Request. Set to POST. URL: https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent?key={{$credentials.apiKey}}. Pause video or grab from description if needed.

Auth: Generic > API Key. Header name: x-goog-api-key. Value: your key. Body: JSON. Paste this structure:

{
“contents”: [
{
“parts”: [
{
“text”: “{{$json.prompt}}”
}
]
}
],
“generationConfig”: {
“responseModalities”: [“PRODUCE_PART_IMAGE”],
“aspectRatio”: “1:1”,
“imageSize”: “1024×1024”
}
}

Prompt pulls from form. Set imageSize to “2048×2048” for 2K or “4096×4096” for 4K.

Mastering Generation Configuration (Resolution and Aspect Ratio)

GenerationConfig rules it all. “responseModalities” turns on images. AspectRatio: 1:1, 16:9, etc. ImageSize: 1K (1024×1024), 2K, or 4K.

Test with “A woman holding a coffee cup with a coffee shop in the background. She’s wearing a dress and a hat.” Execute. Pro photo pops out. Text on signs? Crystal clear. Details pop.

Handling the Output: Converting Base64 to Viewable Files

API sends Base64 string. No image yet. Add “Convert to File” node. Pick “Move Base64 String to a File.”

Drop data from prior node here. Output stays “data.” Run. Now view, download, use the image anywhere.

Section 4: Dynamic Image Manipulation: Editing and Inpainting

Setting Up the Edit Image Workflow

Copy your form. Rename to “Edit Image.” Add file field: “image,” single file, required. Title: “Edit Image.”

Test: Upload dog pic. Prompt: “The dog swims in a pool.” Or celeb shots—it nails them.

Copy generate node too. URL and auth stay same.

Pre-processing: Converting the Input Image to Base64

Before generate, add “Extract from File.” Pick “New File to Base64 String.” Binary field: “image.” Output: “data.”

Run. Gets Base64 ready. Opposite of output convert.

Modifying the API Call for Image Editing

In JSON body, add to parts after text:

{
“inlineData”: {
“mimeType”: “image/png”,
“data”: “{{$json.data}}”
}
}

Text from form prompt now. Image Base64 in data. Generate. Convert output Base64 back to file.

Example: Start with coffee woman. Prompt: “Replace coffee cup with ice cream, background ocean beach.” Same character—necklace, dress, face match. Magic.

Try YouTube thumb: Upload edited pic. “Create YouTube thumbnail with text ‘My best vacation’.” Text reads perfect. Ready to use.

Section 5: Advanced Composition – Combining Multiple Input Images

Structuring the Multi-Image Form Input

Copy edit setup. Rename “Multiple Images.” Keep prompt and image_one (file). Add image_two (file, required).

Up to 12 images work. Two for now. Use case: Character + hat for product promo.

Upload woman, pink hat. Prompt: “Woman wears the pink hat.”

Parallel Base64 Conversion for Multiple Inputs

Copy Base64 nodes. One for image_one (output “image_one”). One for image_two (output “image_two”).

Connect forms to each. Run steps. Both convert clean—no overwrites.

Building the Composite Generation Request

In generate JSON, parts array: text prompt first. Then inlineData objects for each.

Like:

{
“inlineData”: {
“mimeType”: “image/png”,
“data”: “{{$json.image_one}}”
}
},
{
“inlineData”: {
“mimeType”: “image/png”,
“data”: “{{$json.image_two}}”
}
}

Copy-paste blocks easy. Generate. Convert final Base64. Result: Woman rocks pink hat perfectly. Character consistent, product shines.

Conclusion: The Power of Nano Banana Pro in Automated Workflows

You now master Nano Banana Pro in N8N. From basic gens to 4K edits and multi-image mixes. HTTP requests unlock it all—forget node limits.

Key wins: Same characters across edits. Readable text. Product shots that sell. Automate for thumbnails, ads, more.

Build your workflows today—what will you create first?