Nano Banana 2 Workflow for Generating Diverse Bottle Size Variations

Defining the Scaling Challenge in Product Visualization

When developing a new product line, visualizing how a single design translates across different container volumes is critical. Brands often need to present travel-sized, standard, and large-format options simultaneously to stakeholders or for marketing mockups. This process requires maintaining consistent branding elements while accurately altering proportions. The Nano Banana 2 tool facilitates this by allowing users to generate diverse bottle size variations from a single concept. It is important to clarify that Nano Banana refers to the AI image generation and editing tool itself, not a skincare brand, bottle, jar, or any physical subject. Example products generated are generic and unbranded.

The core challenge lies in instructing the model to preserve the aesthetic identity of the bottle—such as its shape, color palette, and texture—while strictly modifying its dimensions. Unlike simple resizing, which distorts the object, this workflow aims to create distinct, realistic instances of the same design family. By leveraging text-to-image capabilities, users can explore multiple scale options without the cost of physical prototyping. However, prompt instructions describe desired outcomes; they do not guarantee identity, label, object, or typography preservation. Users must treat the output as a conceptual visualization rather than a final production-ready asset.

Step-by-Step Input and Prompt Construction

To begin this workflow, you will need a clear description of your base bottle design. Start by defining the material (e.g., frosted glass, matte plastic), the cap style, and the overall silhouette. Your input should be specific about the visual attributes you wish to keep constant across all variations. For instance, if the bottle has a unique curve on the shoulder, ensure this detail is explicitly mentioned in your initial prompt.

Next, construct a usable prompt that targets the three specific sizes: travel, standard, and large. You can use the following example structure to guide your input. Note that these are examples of how to phrase requests and do not guarantee specific results.

Example Prompt: "Generate three distinct bottles side-by-side. On the left, a small travel-sized bottle with a flip-top cap. In the center, a standard full-size bottle with a pump dispenser. On the right, a large luxury-sized bottle with a wide mouth. All bottles must share the same frosted blue glass texture, identical gold foil logo placement, and matching minimalist typography style. The lighting should be soft studio lighting with a white background."

If you have an existing reference image of your base design, you can utilize the image-to-image workflow. Upload the image of your standard bottle and combine it with a text prompt specifying the need for size variations. Be aware that Google describes Nano Banana 2 Lite as focused on speed and cost. It is not optimized for multiple reference inputs or multi-turn sequential editing. Do not recommend it for those workflows without explaining this limitation. For complex scaling tasks involving multiple references, the standard Nano Banana 2 model is more appropriate.

Checkpoints and Iterative Refinement

Once the initial generation is complete, review the outputs against your design requirements. The first checkpoint is proportionality. Ensure the travel bottle looks appropriately compact and the large bottle appears substantial, without looking like a stretched version of the original. The second checkpoint is consistency. Verify that the texture, color, and cap details remain uniform across all three sizes. If the labels appear distorted or the caps look mismatched, this indicates a need for refinement.

Use the iterative nature of the tool to adjust your prompts. If the sizes are too similar, add descriptors like "significantly smaller" or "oversized." If the textures vary, reiterate the material description. Remember that prompt instructions describe desired outcomes; they do not guarantee identity, label, object, or typography preservation. You may need to run several generations to achieve a set where all three bottles look cohesive yet distinct in scale. This process helps identify potential design flaws early, such as a cap that might not fit a larger neck diameter, before any physical molds are created.

Exporting and Using the Visualizations

After achieving a satisfactory result, you can proceed to export the images for your presentation or internal review. These visuals serve as powerful tools for communicating product line scaling strategies to teams and clients. They allow for rapid comparison of how a design feels at different price points or usage scenarios. While the images provide excellent conceptual clarity, they are not a substitute for engineering validation. Always verify functional aspects like ergonomics and packaging logistics through traditional methods.

For users seeking to explore more advanced features or different model capabilities, you can visit the Try Nano Banana page. This platform supports text-to-image and image-to-image workflows, offering a flexible environment for creative experimentation. By following this structured approach, you can efficiently generate diverse bottle size variations, streamlining the early stages of product development and reducing reliance on costly physical prototypes.