Nano Banana Tutorial: Compositing Glazed Pots with Raw Clay Fragments

Understanding the Narrative Goal

Creating a visual story that captures the transformation of clay from its raw state to a finished, glazed pot requires more than just generating a single image. The goal here is compositing stages. You want to show the evolution of the creation process within one cohesive frame. This involves blending textures of rough, unglazed earth with the smooth, reflective surfaces of a completed ceramic piece. By using Nano Banana, you can guide the AI to merge these distinct production phases into a unified artistic composition.

It is important to clarify that Nano Banana refers to the AI image generation and editing tool described in this documentation. It is not a skincare brand, bottle, jar, or physical subject itself. When we discuss the "pot" or "clay," we are referring to the visual content generated by the tool, not a specific commercial product line. This tutorial focuses on the text-to-image and image-to-image workflows available on the platform to achieve this specific artistic intent.

Setting Up Your Workflow for Multi-Stage Blending

To successfully composite different stages of pottery making, you need to select the right model capabilities. Google documents Nano Banana 2 as Gemini 3.1 Flash Image, while Nano Banana Pro corresponds to Gemini 3 Pro Image. For complex tasks involving multiple reference inputs or detailed sequential editing, the Lite version (Gemini 3.1 Flash Lite) has specific limitations. Google describes Nano Banana 2 Lite as focused on speed and cost, noting it is not optimized for multiple reference inputs or multi-turn sequential editing. Therefore, for a task requiring the precise blending of raw and finished elements, relying on the standard Nano Banana 2 or Pro models is advisable over the Lite version unless you accept potential constraints in handling complex references.

Before writing your prompt, consider the structure of your request. The prompt library offers example prompts that users can copy or take into the generator. However, remember that prompt instructions describe desired outcomes; they do not guarantee identity, label, object, or typography preservation. Your success depends on clearly describing the visual relationship between the two states rather than expecting the AI to perfectly replicate specific real-world objects without guidance.

Step-by-Step Guide to Generating the Composite Shot

Follow these numbered steps to create your narrative shot using the tool:

1. Access the Generator: Navigate to the main interface where text-to-image and image-to-image workflows are supported. You can explore the features further at Try Nano Banana.
2. Define the Scene: In your prompt box, explicitly state the desire to see both raw and finished states. Use descriptive language such as "a split composition showing rough unglazed clay fragments on the left and a fully glazed, shiny ceramic pot on the right, seamless transition in the center."
3. Refine Textures: Add details about lighting and material properties. Specify "soft studio lighting highlighting the matte texture of the raw clay and the glossy reflection on the glazed surface."
4. Iterate with Image Inputs: If using the image-to-image workflow, upload a reference of raw clay and another of a glazed pot if the model supports multiple inputs. Note that if you attempt this with the Lite version, results may vary due to its lack of optimization for multiple reference inputs.
5. Review and Adjust: Generate the image and evaluate the blend. If the transition looks jarring, refine the prompt to emphasize "smooth gradient" or "artistic fusion" between the two materials.

Evaluating Results and Troubleshooting Common Issues

Judging the success of your composite shot involves checking the narrative flow. Does the image tell the story of creation? Are the textures distinct yet harmonious? Since prompt instructions do not guarantee object preservation, you might find the AI interprets "fragments" differently each time. If the result lacks the raw clay feel, try adding keywords like "earthy," "unpolished," or "rough." If the glazed pot looks too dull, specify "high gloss," "reflective glaze," or "vibrant finish."

If the image fails to blend the stages correctly, it may be due to the model's interpretation of conflicting concepts. In such cases, simplifying the prompt to focus on one dominant theme before introducing the second element can help. Alternatively, ensure you are not using the Lite version for this specific multi-stage task, as its design prioritizes speed over complex compositional logic. Remember, these examples are illustrative of the tool's capabilities and do not guarantee identical outcomes in every generation.

By following these guidelines, you can effectively use Nano Banana to visualize the journey of pottery making, turning raw materials into a compelling visual narrative.