Gemini Nano: Banana Trend Face Editing Prompts 2025

Lisa Ernst · 16.11.2025 · Technology · 8 min

Nano Banana, the image model behind Google's Gemini, allows the transformation of selfies into collectible figures, retro portraits, or festive looks. This article explains how Nano Banana works, guides you step-by-step through safe face editing, offers 15 prompts to try directly, and discusses privacy and platform rules.

Source: Self-Representation

Introduction & Basics

Gemini is Googles KI-Plattform, the chat, search, and creative tools combined. Nano Banana is the unofficial name for Gemini 2.5 Flash Image, , an image model from Google DeepMind that generates and edits images based on text prompts. According to Modellbeschreibung , Nano Banana was developed as a text-driven image model capable of creating new images and modifying uploaded photos, including fine, local edits.

In practice, this means Gemini Nano Banana Face-Editing: You upload a selfie, describe in a prompt how the scene should look, and the model adjusts clothing, background, lighting, or style without changing your fundamental facial expressions and features. The option to combine multiple images allows you to merge your face with a separately photographed scene, such as a desk or a festival stage.

Nano Banana was built to recognize characters consistently across multiple images, allowing the same person to be placed in different scenes consistently. This consistency is the basis for typical face edits like mini-action figures, 80s portraits, or dramatic movie posters that still clearly resemble you.

Analysis & Context

Nano Banana first appeared anonymously on Bewertungsplattform LMArena , where the community tests image models with real prompts. Google later confirmed that this model is officially Gemini 2.5 Flash Image and made it accessible to the general public on August 26, 2025.

Since then, Nano Banana has been directly integrated into the Gemini-App integriert: You log in with your Google account, select the image function in the web or mobile interface, upload a photo, and edit it with text prompts. Google reports that the image features based on this technology have already reached over 200 Millionen Bearbeitungen ausgelöst and brought more than 10 million new users to the Gemini app.

In parallel, Nano Banana has been gradually integrated into other Google services. For example, in Google Photos, you can open a photo, tap on "Help me edit" tippen , and type in descriptions like "blur the background" or "remove sunglasses," which the model will implement. Media reports describe that Photos even accesses faces from your own library to realistically maintain edits like open eyes or gentler smiles.

At the same time, there are concrete examples of how people are using Nano Banana for selfies. Articles from India show how simple self-portraits become glamouröse Saree-Bilder im Stil von 90er-Jahre-Bollywood-Filmen , with elaborate fabrics, lighting reflections, and vintage backgrounds. Simultaneously, the same articles report on users who found details like an unexpected mole in the AI image irritating and wonder how deeply the model "thinks" into their appearance.

From a safety perspective, Google highlights that all images generated or edited with Nano Banana are tagged with an invisible digital watermark called SynthID to make them identifiable as AI images. While this does not completely prevent misuse, it increases traceability.

The video provides a quick visual overview of how to access Nano Banana in the Gemini interface, upload images, and try out initial edits.

The clip shows additional examples of how to structure prompts for portraits and how small changes in wording can affect the outcome.

Application & Prompts

Let's now look specifically at how you can use Nano Banana for face editing without getting yourself or others into trouble.

Step-by-Step: Preparing Your Selfie

First, the most important principle: only edit your own face or the photos of people who have given you their explicit consent. This is especially important if you plan to publish or share the result later.

Taking a Selfie

Take a photo in even, as natural light as possible, preferably front-on or slightly to the side, so your face is clearly visible. A clean background helps the model focus on you and later integrate a new background cleanly.

Accessing Nano Banana

If you are using the Gemini web interface or the app, log in with your Google account, select "Create image" or a similar option, and upload your selfie. In Google Photos, open the image and tap on "Help me edit", to start a Nano Banana-based edit.

Writing a Prompt

Formulate your request as concretely as possible: What should change (clothing, background, lighting, style), and what should definitely remain the same (your face, your hairstyle, your harmless posture). If something doesn't work out on the first try, describe in the next prompt what gets you closer to your goal, for example, "Keep my face, but make the clothing brighter and the background blurry."

Selecting and Refining a Variant

Nano Banana usually generates multiple suggestions. Choose the version that is closest to your idea and refine it with follow-up prompts, for example, by adjusting colors, light, or details.

15 Prompts for Gemini Nano Banana Face-Editing

The following prompts are formulated to focus on your own face. You can replace individual terms, but stick to the basic idea: no real other people without consent, no sensitive or demeaning scenes.

ACTION FIGURE AND GAMING LOOKS

Prompt 1: Mini Action Figure on a Desk

"Transform my uploaded selfie into a hyper-realistic 3D action figure of myself, original size around 15 centimeters, in a transparent collector's box on a tidy desk, soft studio lighting, clear product photography, my face clearly recognizable."

Prompt 2: Sci-Fi Gaming Avatar

"Stylize my face into a futuristic gaming avatar in sci-fi armor, neon blue and neon magenta glowing details, dark background with holographic interface elements, cinematic lighting, my face remains realistic and friendly."

Prompt 3: Pixel Game Cover

"Create a 16-bit game cover in retro pixel style with my face as the main character in the foreground, simple geometric shapes, limited color palette, clearly legible facial features, stylized city at night in the background."

RETRO 80s/90s

Prompt 4: 80s Glam Portrait

"Transform my selfie into a classic 80s studio portrait, soft pink-blue neon lighting, slight fog, voluminous 80s-style hair, glossy makeup, my face remains natural, geometric neon shapes in the background."

Prompt 5: 90s VHS Screenshot

"Edit my face into a screenshot from a 90s camcorder video, slightly grainy image, VHS timestamp in the lower left, living room atmosphere, warm colors, my facial features remain clear, slight motion blur."

Prompt 6: Polaroid Wall

"Create a single Polaroid photo of myself with a white border, slightly yellowed look, handwritten name below the photo, soft window light, blurred background, my face clearly in focus."

BOLLYWOOD AND DRAMA

Prompt 7: Saree Portrait with Film Lighting

"Stage me in a detailed, colorful saree with golden accents, my face remains distinctly mine, warm cinematic lighting, blurred vintage background, slight depth of field, expression calm and dignified."

Prompt 8: Dramatic Movie Poster

"Design a dramatic movie poster from my selfie, my face large in the foreground, strong color contrast between shadow and light, subtle text areas without readable text, background of an implied city skyline at night."

Prompt 9: Dance Scene in Studio

"Place me as the central figure in a dance scene in a modern studio, traditional festive clothing, dynamic pose, frozen movement, colored spotlights, my face stays sharper than the background."

FESTIVE LOOKS: DIWALI, CHRISTMAS, AND FESTIVALS

Prompt 10: Sea of Lights

"Transform my selfie into a portrait at night with many small string lights in the background, bokeh-like light points, warm gold tones, my face from the front, soft side lighting, calm, happy expression."

Prompt 11: Winter Sweater Portrait

"Show me in a thick knitted winter sweater with a subtle pattern, light snow in the background, soft, cool daylight, my face in half-profile, breath visible in the cold air."

Prompt 12: Festival Stage

"Stage me as a performer on an open-air stage, my face in the foreground, blurred audience and colorful lights in the background, light fog, energetic mood, modern clothing, but no recognizable brand logos."

BONUS: HARYANA-STYLE COLOR BLOCKS

The so-called Haryana look is often described with warm, slightly desaturated colors, soft contrast, and a very calm, documentary mood. Use it as a color mood, not as a template for a specific person.

Prompt 13: Haryana-Style Portrait

"Edit my selfie into a calm portrait in the Haryana color look: warm, slightly desaturated colors, soft light, subtle contrast, blurred landscape background, my face sharp and natural."

Prompt 14: Haryana Street Frame

"Stage me in a daylight street scene, color mood like in North Indian documentary photos, warm tones, fine dust in the air, flat light, my face remains clear and friendly."

Prompt 15: Haryana DC-Inspired Color Look

"Keep my current outfit and pose, but apply a color look reminiscent of official portraits from North Indian administrative buildings: warm beige and brown tones, subtle vignette, clean background wall, my face remains unchanged."

Tips for Posts and Workflow

If you want to share results, it helps to plan the image composition and format. On many platforms, square images or vertical formats between 4:5 and 9:16 work particularly well because they take up more space on smartphones. So, you can even request in the prompt that your portrait be centrally placed and not cropped too much.

Before-and-after collages can be easily created in an app like CapCut by placing the original and the AI version next to or on top of each other. You can then adjust color looks later and work with text without having to further modify the AI image itself. For hashtags and captions, be careful not to present your image as a real documentary snapshot when it is clearly an AI-generated image.

It is also important to know the rules of the platform where you want to post. Many services prohibit, for example, explizite Inhalte, violent scenes or misleading representations of real people, even if they would be technically allowed. Stick to simple principles: no edits that could harm others, no deepfakes without consent, and no imitation of official documents like IDs or identity cards.

Reactions & Open Questions

Many tech media outlets emphasize the creative possibilities. Tutorials show step-by-step how to find Nano Banana in the Gemini app, upload a selfie, try out initial prompts, and become increasingly precise with repetitions. Blogs from prompt communities collect lists of ideas on how to represent yourself as a collectible figure, a character on a magazine cover, or in the style of analog photos.

On the other hand, data privacy and security experts warn against too casual use of selfies. Articles from India and international security blogs highlight that edited images can be misused for Deepfakes, Identitätsdiebstahl oder Social-Engineering-Angriffe if they fall into the wrong hands. Particularly sensitive is the use on third-party sites that pretend to be official access points to Nano Banana but have very opaque data protection policies themselves.

Google itself tries to emphasize a balance in official statements: the company refers to SynthID, internal security checks and to the fact that models should not simply generate arbitrary people in problematic contexts. At the same time, external experts maintain that technical safeguards alone are not enough and users must consciously decide which images they upload and where.

Even though Nano Banana is technically impressive, some points remain open. It is not yet transparent in detail how long original photos and generated images are stored in different Google services and how granularly you can remove them from training and evaluation processes. Security blogs also ask how robust SynthID markings are against later post-processing or aggressive compression and how easily external tools can reliably detect AI images.

From a regulatory perspective, questions arise about how AI images should be evaluated in media, advertising, or legal proceedings. If an edited image looks very real but has an invisible watermark, clearer standards are needed on how platforms and institutions should handle it. In parallel, researchers are investigating how AI-assisted face editing affects self-image, body perception, and social norms in the long term, similar to how beauty filters did in the past.

For you, this means: the technology is relatively easily accessible, but the social and legal responses are not yet. The more consciously you decide now which images you create, save, and share, the easier it will be later to understand what you want to be responsible for and what you don't.

Conclusion & Recommendations

Gemini Nano Banana face editing can be a creative tool that allows you to put your own face into new roles, stories, and looks without hiding behind anonymous filters. If you use the technology with some clear rules of engagement, it is more of an amplifier for imagination than a risk: only use images with consent, refrain from manipulative deepfakes, check privacy settings, and honestly label AI images where appropriate.

The 15 prompts above are a starting point, not a framework to restrict you. If you understand how Nano Banana works with light, style, background, and composition, you can gradually formulate your own ideas and always check if the result still feels right. This way, an AI function becomes a tool that you consciously control – not the other way around.