Nano Banana (Gemini Image Generation)

Nano Banana provides adapters for Google's Gemini image generation models, enabling text-to-image generation, image editing, multi-image composition, batch generation, and (via NanoBanana2LiteAdapter) a lightweight virtual try-on convenience method.

Overview

The tryon.api.nano_banana module provides four adapters:

• NanoBananaAdapter: Gemini 2.5 Flash Image — Fast, efficient, 1024px resolution
• NanoBananaProAdapter: Gemini 3 Pro Image Preview — Advanced, up to 4K resolution, search grounding
• NanoBanana2Adapter: Gemini 3.1 Flash Image (Nano Banana 2) — Pro capabilities at Flash speed; 512px–4K, subject consistency, precise instruction following. See Google's announcement.
• NanoBanana2LiteAdapter: Gemini 3.1 Flash-Lite Image (Nano Banana 2 Lite) — Google's fastest/cheapest tier; 1K resolution only, up to 14 reference images, but "not optimized for multiple reference inputs or multi-turn sequential editing" per Google's docs. See Google DeepMind's announcement.

Prerequisites

1. Google Gemini API Key:

   • Sign up at Google AI Studio
   • Get your API key from API Keys page
   • Set GEMINI_API_KEY environment variable

2. Install Dependencies:

   pip install google-genai

NanoBananaAdapter (Gemini 2.5 Flash Image)

Fast and efficient image generation optimized for high-volume, low-latency tasks.

Initialization

from tryon.api.nano_banana import NanoBananaAdapter

# Using environment variable
adapter = NanoBananaAdapter()

# Or specify API key directly
adapter = NanoBananaAdapter(api_key="your_api_key")

Text-to-Image Generation

Generate images from text descriptions.

images = adapter.generate_text_to_image(
    prompt="A stylish fashion model wearing a modern casual outfit in a studio setting",
    aspect_ratio="16:9"  # Optional
)

# Save results
for idx, image in enumerate(images):
    image.save(f"result_{idx}.png")

Parameters:

• prompt (str): Text description of the image to generate
• aspect_ratio (str, optional): Aspect ratio. Options: "1:1", "2:3", "3:2", "3:4", "4:3", "4:5", "5:4", "9:16", "16:9", "21:9"

Returns: List[Image.Image] - List of PIL Image objects

Image Editing

Edit images using text prompts to add, remove, or modify elements.

images = adapter.generate_image_edit(
    image="person.jpg",
    prompt="Change the outfit to a formal business suit",
    aspect_ratio="16:9"  # Optional
)

Parameters:

• image (str/PIL.Image): Input image (file path, URL, PIL Image, or base64)
• prompt (str): Text description of edits to make
• aspect_ratio (str, optional): Aspect ratio for output

Returns: List[Image.Image] - List of edited PIL Image objects

Multi-Image Composition

Combine multiple images with style transfer and composition.

images = adapter.generate_multi_image(
    images=["outfit1.jpg", "outfit2.jpg"],
    prompt="Create a fashion catalog layout combining these clothing styles",
    aspect_ratio="16:9"  # Optional
)

Parameters:

• images (List[str/PIL.Image]): List of input images
• prompt (str): Text description of how to combine images
• aspect_ratio (str, optional): Aspect ratio for output

Returns: List[Image.Image] - List of composed PIL Image objects

Batch Generation

Generate multiple images from multiple prompts.

results = adapter.generate_batch(
    prompts=[
        "A fashion model showcasing summer collection",
        "Professional photography of formal wear",
        "Casual street style outfit on a model"
    ],
    aspect_ratio="16:9"  # Optional
)

# Save all results
for prompt_idx, images in enumerate(results):
    for img_idx, image in enumerate(images):
        image.save(f"batch_{prompt_idx}_{img_idx}.png")

Parameters:

• prompts (List[str]): List of text prompts
• aspect_ratio (str, optional): Aspect ratio for all images

Returns: List[List[Image.Image]] - List of lists, where each inner list contains images for that prompt

Supported Aspect Ratios

Aspect Ratio	Resolution	Tokens
1:1	1024x1024	1290
2:3	832x1248	1290
3:2	1248x832	1290
3:4	864x1184	1290
4:3	1184x864	1290
4:5	896x1152	1290
5:4	1152x896	1290
9:16	768x1344	1290
16:9	1344x768	1290
21:9	1536x672	1290

NanoBananaProAdapter (Gemini 3 Pro Image Preview)

Advanced image generation with 4K resolution support and search grounding.

Initialization

from tryon.api.nano_banana import NanoBananaProAdapter

# Using environment variable
adapter = NanoBananaProAdapter()

# Or specify API key directly
adapter = NanoBananaProAdapter(api_key="your_api_key")

Text-to-Image Generation

Generate images with 1K, 2K, or 4K resolution.

images = adapter.generate_text_to_image(
    prompt="Professional fashion photography of elegant evening wear on a runway",
    resolution="4K",  # Options: "1K", "2K", "4K"
    aspect_ratio="16:9",
    use_search_grounding=True  # Optional: Use Google Search for real-world grounding
)

Parameters:

• prompt (str): Text description of the image to generate
• resolution (str): Resolution level. Options: "1K", "2K", "4K". Default: "1K"
• aspect_ratio (str, optional): Aspect ratio (same options as Nano Banana)
• use_search_grounding (bool, optional): Use Google Search for real-world grounding. Default: False

Returns: List[Image.Image] - List of PIL Image objects

Image Editing

Edit images with high-resolution output.

images = adapter.generate_image_edit(
    image="person.jpg",
    prompt="Change the outfit to a formal business suit",
    resolution="2K",  # Options: "1K", "2K", "4K"
    aspect_ratio="16:9"
)

Parameters:

• image (str/PIL.Image): Input image
• prompt (str): Text description of edits to make
• resolution (str): Resolution level. Options: "1K", "2K", "4K". Default: "1K"
• aspect_ratio (str, optional): Aspect ratio for output

Returns: List[Image.Image] - List of edited PIL Image objects

Multi-Image Composition

Compose multiple images with high-resolution output.

images = adapter.generate_multi_image(
    images=["outfit1.jpg", "outfit2.jpg"],
    prompt="Create a fashion catalog layout combining these clothing styles",
    resolution="2K",
    aspect_ratio="16:9"
)

Parameters:

• images (List[str/PIL.Image]): List of input images
• prompt (str): Text description of how to combine images
• resolution (str): Resolution level. Options: "1K", "2K", "4K". Default: "1K"
• aspect_ratio (str, optional): Aspect ratio for output

Returns: List[Image.Image] - List of composed PIL Image objects

Batch Generation

Generate multiple images in batch with high resolution.

results = adapter.generate_batch(
    prompts=[
        "A fashion model showcasing summer collection",
        "Professional photography of formal wear",
        "Casual street style outfit on a model"
    ],
    resolution="2K",
    aspect_ratio="16:9"
)

Parameters:

• prompts (List[str]): List of text prompts
• resolution (str): Resolution level for all images. Options: "1K", "2K", "4K". Default: "1K"
• aspect_ratio (str, optional): Aspect ratio for all images

Returns: List[List[Image.Image]] - List of lists, where each inner list contains images for that prompt

Supported Resolutions and Aspect Ratios

Nano Banana Pro supports the same 10 aspect ratios as Nano Banana, but with resolution-specific dimensions:

Aspect Ratio	1K Resolution	2K Resolution	4K Resolution
1:1	1024x1024	2048x2048	4096x4096
16:9	1376x768	2752x1536	5504x3072
9:16	768x1376	1536x2752	3072x5504
...	...	...	...

See the Gemini Image Generation Documentation for complete resolution tables.

NanoBanana2Adapter (Gemini 3.1 Flash Image — Nano Banana 2)

Nano Banana 2 combines the advanced capabilities of Nano Banana Pro with the speed of Gemini Flash: production-ready specs (512px–4K), subject consistency, and precise instruction following. Ideal for rapid iteration and production assets.

Reference: Nano Banana 2: Google's latest AI image generation model

Initialization

from tryon.api.nano_banana import NanoBanana2Adapter

adapter = NanoBanana2Adapter()
# Or: adapter = NanoBanana2Adapter(api_key="your_api_key")

Text-to-Image Generation

images = adapter.generate_text_to_image(
    prompt="A fashion model wearing seasonal collection",
    resolution="2K",   # "1K", "2K", or "4K". Default: "2K"
    aspect_ratio="16:9",
    use_search_grounding=False  # Optional
)

Parameters: Same as Nano Banana Pro (prompt, resolution, aspect_ratio, use_search_grounding). Default resolution is "2K".

Returns: List[Image.Image]

Image Editing and Multi-Image Composition

Same method signatures as Nano Banana Pro: generate_image_edit(), generate_multi_image(), generate_batch(), all with resolution (default "2K"), aspect_ratio, and optional use_search_grounding.

When to Use Which Model

Use case	Adapter
Fastest iteration, 1024px	NanoBananaAdapter
Maximum quality, 4K, search grounding	NanoBananaProAdapter
Pro quality at Flash speed, rapid edits	NanoBanana2Adapter
Lowest latency/cost, high-volume pipelines	NanoBanana2LiteAdapter

NanoBanana2LiteAdapter (Gemini 3.1 Flash-Lite Image — Nano Banana 2 Lite)

Google's fastest and cheapest Gemini image model, engineered for velocity and scale where speed and cost are the primary operational constraints. Capped at 1K resolution; per Google's docs it is "not optimized for multiple reference inputs or multi-turn sequential editing" -- use NanoBanana2Adapter instead when reference-image fidelity matters more than latency/cost.

Reference: Gemini 3.1 Flash-Lite Image — Google DeepMind

Initialization

from tryon.api.nano_banana import NanoBanana2LiteAdapter

adapter = NanoBanana2LiteAdapter()
# Or: adapter = NanoBanana2LiteAdapter(api_key="your_api_key")

Text-to-Image Generation

images = adapter.generate_text_to_image(
    prompt="A fashion model wearing a summer collection",
    aspect_ratio="16:9",  # Optional
)

Parameters:

• prompt (str): Text description of the image to generate
• aspect_ratio (str, optional): Aspect ratio. Options: "1:1", "2:3", "3:2", "3:4", "4:3", "4:5", "5:4", "9:16", "16:9", "21:9"

Returns: List[Image.Image] — always at 1K resolution (no resolution parameter)

Image Editing and Multi-Image Composition

Same method signatures as the other adapters: generate_image_edit(image, prompt, aspect_ratio, ...), generate_multi_image(images, prompt, aspect_ratio, ...), generate_batch(prompts, aspect_ratio, ...). None of them accept a resolution or use_search_grounding parameter, since Flash-Lite Image only supports 1K output and doesn't support search grounding.

Virtual Try-On (Multi-Image Composition)

NanoBanana2LiteAdapter adds a generate_virtual_tryon() convenience method that composes a garment onto a person image via generate_multi_image(), with a sensible default styling prompt (mirroring FluxVTONAdapter.generate). This is a fast/cheap option, not the highest-fidelity one -- prefer a dedicated VTON model (FLUX VTO, Nova Canvas, Kling AI, Segmind, or P-Image-Try-On) when garment-fit accuracy matters more.

images = adapter.generate_virtual_tryon(
    person="person.jpg",
    garment="jacket.jpg",
    garment_description="olive green bomber jacket",  # Optional, builds the default prompt
    # or: prompt="Full custom styling instruction"     # Optional, overrides garment_description
)
images[0].save("result.png")

Parameters:

• person / source_image / person_image / model_image (aliases): Person/model image
• garment / reference_image / garment_image / cloth_image (aliases): Garment reference image
• prompt (str, optional): Full styling prompt override
• garment_description (str, optional): Short garment description used to build the default prompt
• aspect_ratio (str, optional): Aspect ratio for the output

Returns: List[Image.Image]

Also available from the opentryon CLI / MCP server as vton --model nano-banana-2-lite.

This is composition try-on via the Gemini Developer API (GEMINI_API_KEY). Dedicated Google Cloud Virtual Try-On is --model google-vton (Vertex virtual-try-on-001, ADC + GOOGLE_CLOUD_PROJECT). See Google Virtual Try-On.

Command Line Usage

Use the image_gen.py script for command-line image generation:

# Text-to-image with Nano Banana
python image_gen.py --provider nano-banana --prompt "A stylish fashion model wearing a modern casual outfit"

# Text-to-image with Nano Banana Pro (4K)
python image_gen.py --provider nano-banana-pro --prompt "Professional fashion photography of elegant evening wear" --resolution 4K

# Text-to-image with Nano Banana 2 (Pro quality at Flash speed)
python image_gen.py --provider nano-banana-2 --prompt "A fashion model wearing seasonal collection" --resolution 2K

# Image editing
python image_gen.py --provider nano-banana --mode edit --image person.jpg --prompt "Change the outfit to a formal business suit"

# Multi-image composition
python image_gen.py --provider nano-banana --mode compose --images outfit1.jpg outfit2.jpg --prompt "Create a fashion catalog layout"

# Batch generation
python image_gen.py --provider nano-banana --batch prompts.txt --output-dir results/

Input Format Support

Both adapters support multiple input formats:

• File paths: "path/to/image.jpg"
• URLs: "https://example.com/image.jpg"
• PIL Images: Image.open("image.jpg")
• File-like objects: io.BytesIO(image_bytes)
• Base64 strings: Base64-encoded image data

Error Handling

from tryon.api.nano_banana import NanoBananaAdapter

try:
    adapter = NanoBananaAdapter()
    images = adapter.generate_text_to_image("A fashion model showcasing seasonal clothing")
except ValueError as e:
    # Validation errors (missing API key, invalid parameters, etc.)
    print(f"Validation error: {e}")
except ImportError as e:
    # Missing dependencies
    print(f"Import error: {e}. Install google-genai: pip install google-genai")
except Exception as e:
    # API errors, network errors, etc.
    print(f"API error: {e}")

Best Practices

1. Use Nano Banana for: High-volume, low-latency tasks, fast iteration at 1024px
2. Use Nano Banana Pro for: Professional production, 4K resolution needs, search grounding
3. Use Nano Banana 2 for: Pro-level quality with Flash speed, rapid edits, subject consistency, 1K/2K/4K
4. Cache API Key: Use environment variables instead of hardcoding
5. Batch Processing: Use generate_batch() for multiple prompts to optimize API calls
6. Aspect Ratios: Choose appropriate aspect ratios for your use case
7. Error Handling: Always wrap API calls in try-except blocks

Examples

Complete Workflow

from tryon.api.nano_banana import NanoBananaAdapter, NanoBananaProAdapter, NanoBanana2Adapter

# Fast generation with Nano Banana
fast_adapter = NanoBananaAdapter()
images = fast_adapter.generate_text_to_image(
    prompt="A stylish fashion model wearing a modern casual outfit",
    aspect_ratio="16:9"
)
images[0].save("fast_result.png")

# Pro quality at Flash speed with Nano Banana 2
nb2_adapter = NanoBanana2Adapter()
images = nb2_adapter.generate_text_to_image(
    prompt="A fashion model wearing seasonal collection",
    resolution="2K",
    aspect_ratio="16:9"
)
images[0].save("nb2_result.png")

# High-quality generation with Nano Banana Pro
pro_adapter = NanoBananaProAdapter()
images = pro_adapter.generate_text_to_image(
    prompt="Professional fashion photography of elegant evening wear on a runway",
    resolution="4K",
    aspect_ratio="16:9",
    use_search_grounding=True
)
images[0].save("pro_result.png")

Iterative Refinement

from tryon.api.nano_banana import NanoBananaProAdapter

adapter = NanoBananaProAdapter()

# Initial generation
images = adapter.generate_text_to_image("A fashion model wearing casual street style")
current_image = images[0]

# Refine with editing
images = adapter.generate_image_edit(
    image=current_image,
    prompt="Change to formal evening wear with elegant accessories"
)
refined_image = images[0]
refined_image.save("refined.png")

Reference

• Gemini Image Generation Documentation
• Nano Banana 2 (Gemini 3.1 Flash Image) — Google Blog
• Google AI Studio
• API Keys