Nano Banana 2 Prompt Library: Technical Diagram Labeling Examples

Why Precise Labels Matter in Technical Diagrams

When working with complex schematics, mechanical drawings, or engineering blueprints, clarity is paramount. A single mislabeled component can lead to significant misunderstandings in manufacturing, assembly, or maintenance. The Nano Banana 2 tool offers a robust environment for refining these visuals, specifically through its text-to-image and image-to-image workflows. However, achieving accurate typography and correct placement of labels requires more than just a simple request; it demands structured prompting.

The Nano Banana 2 prompt library provides a curated collection of example prompts that users can copy directly into the generator. These instructions are designed to describe desired outcomes clearly. It is important to remember that while these prompts guide the AI, they do not guarantee identity, label, object, or typography preservation in every instance. Users should treat these as starting points to be refined based on their specific diagram complexity. This article curates five materially different usable prompts from the library, explaining their specific use cases, optimal scenarios, and necessary adjustments.

Prompt Example 1: High-Fidelity Component Identification

Prompt: "Add precise, high-contrast black sans-serif labels to this mechanical diagram. Identify the piston, cylinder head, and connecting rod. Ensure text is legible against the white background and aligned with leader lines pointing to the correct parts."

Use Case: This prompt is ideal for standardizing existing diagrams where the visual elements are clear but lack annotation. It focuses on adding new text without altering the underlying drawing geometry.

When to Use: Utilize this when you have a clean line drawing or a rendered 3D model and need to add missing part names quickly. It works best for static images where the orientation of the text does not need to follow a curved surface.

Adjustments: If the generated text appears blurry, increase the emphasis on "high-contrast" and specify the font weight (e.g., "bold sans-serif"). If the leader lines are misplaced, explicitly describe the target coordinates or relative positions, such as "top-right corner" or "center-left." Note that this approach relies on the model's ability to interpret spatial relationships within the image.

Prompt Example 2: Schematic Symbol Annotation

Prompt: "Annotate this electrical schematic with standard ISO symbols and text. Label the resistor as 'R1', capacitor as 'C1', and transistor as 'Q1'. Maintain the original circuit topology and ensure text boxes do not overlap with wire connections."

Use Case: This variation targets electronic schematics where specific nomenclature and symbol standards are critical. It emphasizes maintaining the integrity of the circuit logic while overlaying identifiers.

When to Use: Choose this for circuit boards or wiring diagrams where component values and types must be preserved exactly. It is particularly useful when converting hand-drawn sketches into standardized digital formats.

Adjustments: If the AI alters the circuit lines, reinforce the instruction to "preserve original topology." For overlapping text, add constraints like "keep text outside the bounding box of the wires." Since Google describes Nano Banana 2 Lite as focused on speed and cost, avoid using it for this workflow if multiple reference inputs or multi-turn sequential editing are required, as it is not optimized for those tasks.

Prompt Example 3: Multi-Language Technical Documentation

Prompt: "Translate and overlay technical labels onto this hydraulic system diagram. Change 'Pressure Valve' to 'Válvula de Presión' and 'Flow Meter' to 'Medidor de Flujo'. Keep the font style consistent with the existing English labels and ensure right-to-left alignment if necessary."

Use Case: This prompt addresses international teams requiring localized documentation. It combines translation capabilities with strict layout constraints to match existing design languages.

When to Use: Deploy this when preparing manuals for global distribution or when integrating foreign language components into a unified system. It helps maintain visual consistency across different linguistic versions of the same document.

Adjustments: Be aware that character spacing may vary between languages. You may need to manually tweak the prompt to specify "adjust letter spacing" or "increase padding around text." If the output distorts the diagram, try breaking the task into smaller steps, though be mindful of the limitations of the Lite version regarding sequential editing.

Prompt Example 4: Dynamic Arrow and Leader Line Integration

Prompt: "Generate dynamic leader lines connecting text labels to moving parts in this animation frame. Label the 'Input Shaft' and 'Output Gear'. Ensure the lines are thin, dashed, and terminate with arrowheads at the component edges."

Use Case: This is tailored for animated sequences or frames where components are in motion, requiring labels that track or point dynamically to shifting elements.

When to Use: Use this for creating instructional videos or interactive guides where static labels would be misleading. It ensures that the viewer understands which part is being discussed even during movement.

Adjustments: If the lines appear too thick or solid, specify "thin, dashed lines" again. If the arrowheads are missing, explicitly state "terminate with arrowheads." Remember that prompt instructions describe desired outcomes but do not guarantee object preservation; you may need to iterate several times to get the perfect alignment.

Prompt Example 5: Color-Coded Legend Generation

Prompt: "Create a color-coded legend box for this thermal map diagram. Assign colors to 'Cold Zone', 'Warm Zone', and 'Hot Zone' based on the provided palette. Place the legend in the bottom-right corner with a semi-transparent background."

Use Case: This prompt focuses on generating explanatory legends rather than direct part labeling. It is essential for data visualization where color gradients represent quantitative data.

When to Use: Ideal for scientific reports, weather maps, or stress analysis diagrams where a key is needed to interpret the visual data correctly.

Adjustments: If the legend overlaps the main data, specify "ensure no overlap with primary data areas." If the colors are inaccurate, provide hex codes or specific color names. Always verify the final output against the source data, as the AI interprets descriptions rather than calculating exact values.

For those ready to experiment with these techniques, Try Nano Banana to access the full range of models including Nano Banana 2 and Nano Banana Pro. While these examples serve as a strong foundation, always test prompts with your specific images to achieve the best results. The goal is to enhance clarity, not to replace human verification of technical accuracy.