nano-banana-pro:official is an image generation and editing entry point based on Google Gemini 3 Pro Image, suited for creations with high requirements for text, composition, and asset blending. It can generate new images from text or modify scenes, appearance, and layouts using reference images, helping brand teams create posters, product visuals, and series assets.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Gemini 3 Pro Image is the base model name; the resolutions, aspect ratios, quantities, and delivery methods above are the invocation specifications for this platform entry.
A key strength of Nano Banana Pro is in-image text rendering, making it suitable for headlines, product selling points, and explanatory visuals. When creating, you can explicitly provide copy, placement, and hierarchy so that text participates in the composition together with the main subject, rather than generating only a text-free background. Before formal release, spelling, punctuation, and small text should still be checked character by character.
By combining text and reference images, you can integrate people, products, and environments into a single image, or adjust the background, colors, and composition around existing assets. Prompts should specify the role of each image and which appearances need to be preserved, making this suitable for product compositing and series visual creation with clear subject constraints.
Edit mode is suited to expressing specific changes, such as replacing the background, adjusting lighting, or changing materials. Use the result from the previous round as the next round's reference image, then describe the modification goal to progressively refine the image. Resolution and aspect ratio can be selected independently, making it easy to create square displays, portrait covers, and horizontal hero visuals from the same concept.
Enter the campaign theme, exact copy, brand colors, and composition requirements to generate poster concepts with headlines and key selling points. When existing visuals need to be continued, brand assets can be added as references. First confirm the layout and text, then choose a resolution suitable for delivery to obtain images for review and subsequent layout work.
Provide product images and describe the usage environment, lighting direction, and display arrangement to create home, office, or outdoor scene images. When editing, explicitly require that packaging text, colors, and structure be retained to help focus the scope of changes. Before delivery, compare against the original image to check the product appearance and avoid mistaking generated details for actual product features.
Use people, clothing, props, or backgrounds as different reference assets, describe their relationships within the image, and create content concepts with a consistent theme. Use count to obtain multiple candidates, then continue editing the selected result. This approach is suitable for content planning, outfit concept presentations, and visual exploration of different compositions.
If an image needs not only to look good but also to carry clear headlines, product descriptions, or relationships among multiple reference assets, Nano Banana Pro is better aligned with this type of task. Compared with Nano Banana 2, which is geared toward everyday generation, the focus of choosing Pro is text and compositing requirements, rather than assuming that all tasks receive the same improvement.
The base nano-banana is based on Gemini 2.5 Flash Image and is suitable for rapid content production and prototype validation; Pro is better suited for brand visuals and detailed editing. You can first validate the theme and composition, then use Pro to refine key images. :official is the public invocation ID suffix for this model, not a new Google model version, nor does it mean results are exempt from review.
Choose generate for text-based creation; choose edit to modify existing assets, provide reference images with image_urls, and separately specify what to retain and what to change.
Specify model=nano-banana-pro:official, action, and prompt for /nano-banana/images; start with aspect_ratio=1:1, resolution=2K, and count=1, setting the aspect ratio and resolution separately.
Retrieve images from data[].image_url; for asynchronous requests, query or receive callbacks using task_id. When continuing modifications, pass in the selected image again and narrow the scope of changes in each round.
The first image defines the product, the second image defines the background, and the third image provides the color palette; naturally place the product on the background tabletop, preserve the packaging and label, and generate a vertical advertising visual.
First explain the role of each reference image; check lighting integration, product proportions, and labels, then decide whether to proceed to final layout production.
It is the public invocation ID for Nano Banana Pro, with Google Gemini 3 Pro Image as the underlying model. :official distinguishes the invocation endpoint and should not be understood as a new native version. When using it, enter the full ID and structure requests according to the operations, aspect ratio, and resolution options provided by this endpoint.
Use generate when creating from text; use edit when modifying or compositing around an existing image, and provide reference images through image_urls. Both operations require a prompt. Edit prompts should preferably specify separately what needs to change and which subject, colors, or layout must be retained.
Provide exact copy, and clearly specify the position and visual hierarchy of titles and descriptive text; avoid cramming all information into a very small area. Proofread character by character after generation, and shorten the text or revise in multiple passes if necessary. For designs that must strictly use specified fonts, you can use the generated result as a visual draft and continue typesetting.
The aspect ratio options for this endpoint do not include 4:5; you can choose 1:1, 3:2, 2:3, 16:9, 9:16, 4:3, or 3:4. If delivery requires 4:5, it is recommended to first generate using a similar portrait ratio, reserve edge space in the prompt, and then crop to the target size to avoid cutting off text and the main subject.
Place the URL of the generated image into image_urls in the next editing request, and write specific modification goals. After submitting via POST /nano-banana/images, the result includes a task identifier and image link; batch workflows can combine async and callback_url to receive results after the task is complete.
Choose a development approach or application suited to this model and follow its setup guide.
Python SDK Integration Guide
SDK installation, authentication and examplesTypeScript SDK Integration Guide
SDK installation, authentication and examples
