20 of the best prompts for Nano Banana prompts for photo editing, step by step across 4 stages. Works with ChatGPT, Claude, and Gemini.
20 of the best prompts for Nano Banana prompts for photo editing, step by step across 4 stages. Works with ChatGPT, Claude, and Gemini.
Most people try to use AI for Nano Banana Prompts for Photo Editing with a single vague prompt and get generic results. This guide takes a different approach: 4 targeted stages, from Core editing moves through Creative transformations, each with a prompt that gives the AI exactly the context it needs. Edit real photos with plain-language instructions: remove objects, change backgrounds, fix lighting, restore old photos, and retouch portraits, keeping everything you did not mention untouched. Works with ChatGPT, Claude, and Gemini.
Nano Banana’s signature ability is conversational editing: you upload a photo, describe the change, and it preserves everything else. These prompts teach the instruction patterns that make edits precise instead of destructive.
The precise edit formula
Edit this photo with one change only: [THE CHANGE: REMOVE THE TRASH CAN ON THE LEFT / CHANGE THE WALL COLOR TO SAGE GREEN / CLOSE THE OPEN CABINET DOOR]. Everything else must remain exactly as it is: same people, same lighting, same colors, same framing, same details. Fill any revealed area so it matches the surrounding scene naturally.
Remove distracting objects
Clean up this photo by removing: [LIST OBJECTS: THE POWER LINES CROSSING THE SKY / THE STRANGER IN THE BACKGROUND / THE WATER BOTTLE ON THE TABLE / THE EXIT SIGN]. Reconstruct what sits behind each removed object so the fill is seamless: continuous sky, wall texture, background crowd. Do not smooth, sharpen, or recolor anything else in the image.
Replace the background, keep the subject
Keep the [PERSON / PRODUCT / PET] in this photo exactly as they are: identical pose, expression, clothing, and edge detail including [TRICKY EDGES: HAIR / FUR / GLASS]. Replace only the background with [NEW BACKGROUND: A CLEAN WHITE STUDIO / A BLURRED CITY STREET AT DUSK / A WARM INTERIOR]. Match the subject’s existing lighting direction so they look photographed there, not cut out.
Fix exposure and color
Correct this photo’s technical problems without changing its content: [PROBLEMS: IT IS UNDEREXPOSED AND THE SHADOWS ARE MUDDY / THE WHITE BALANCE IS TOO YELLOW FROM INDOOR LIGHTING / THE SKY IS BLOWN OUT]. Target result: balanced exposure with detail in shadows and highlights, neutral accurate colors, and skin tones that look natural. Keep the edit honest: this should look like the photo was taken correctly, not filtered.
Add something that belongs
Add [OBJECT: A COFFEE CUP ON THE DESK / STRING LIGHTS ALONG THE WALL / A DOG SLEEPING BY THE CHAIR] to this photo. It must inherit the scene’s physics: correct scale for its position, shadows cast in the same direction as existing shadows, lighting and color temperature matching the room, and natural placement that does not overlap awkwardly with existing objects.
Retouching people is the highest-stakes editing: overdone edits look fake instantly. These prompts aim for the professional standard, natural results where the person still looks like themselves.
Natural skin retouch
Retouch this portrait to a natural professional standard: reduce [TEMPORARY ISSUES: THE BLEMISH ON THE CHIN / REDNESS AROUND THE NOSE / SHINE ON THE FOREHEAD] while fully preserving skin texture, pores, freckles, and fine lines. Do not smooth the skin into plastic. The result should look like this person on a good day, not a different person.
Fix closed eyes or expression
This photo is perfect except [ISSUE: MY EYES ARE HALF-CLOSED / ONE PERSON IS MID-BLINK]. Adjust only that: open the eyes naturally, matching this person’s real eye shape and color visible in the image, with a natural gaze direction toward [THE CAMERA / WHERE THEY WERE LOOKING]. Do not alter the smile, face shape, or anything else.
Wardrobe and grooming cleanup
Polish the presentation details in this portrait: [FIXES: FLATTEN THE COLLAR THAT IS STICKING UP / REMOVE THE LINT ON THE BLAZER / TAME THE FLYAWAY HAIRS ON THE RIGHT SIDE / STRAIGHTEN THE NECKLACE]. These are the small fixes a stylist would make on set. Face, expression, pose, and lighting stay untouched.
Update the setting for professional use
I need this casual photo to work as a [LINKEDIN HEADSHOT / TEAM PAGE PHOTO / SPEAKER BIO PHOTO]. Keep my face, expression, and hair exactly as they are. Change the setting to [PROFESSIONAL CONTEXT: A SOFTLY BLURRED MODERN OFFICE / A NEUTRAL STUDIO BACKDROP], adjust my clothing to [ATTIRE: A DARK BLAZER OVER A PLAIN TOP], and light it like a professional headshot: soft key light, gentle background separation.
Group photo repair
Fix this group photo’s problems while keeping every person recognizable: [FIXES: PERSON SECOND FROM LEFT HAS EYES CLOSED, OPEN THEM NATURALLY / REMOVE THE PHOTOBOMBER BEHIND US / BRIGHTEN THE TWO FACES IN SHADOW ON THE RIGHT TO MATCH THE REST]. Nobody’s face shape, body, or position changes. The goal is the shot we thought we took.
Old, damaged, or low-quality photos carry memories no reshoot can replace. These prompts recover them: repairing damage, colorizing, and enhancing quality while respecting what the original actually looked like.
Restore a damaged print
Restore this old photograph: repair the [DAMAGE: CREASES ACROSS THE CORNER / TORN EDGE / WATER STAINS / SCRATCHES AND DUST], reconstructing the damaged areas to match the surrounding image content. Preserve the authentic character: the original grain, tonal range, and period feel stay. This is conservation, not modernization.
Colorize a black and white photo
Colorize this black and white photo with historically plausible colors. Known facts to honor: [DETAILS: THE DRESS WAS PALE BLUE / THIS IS [PLACE] around [YEAR] / the car was dark green]. Skin tones natural and varied, fabrics and environment in period-accurate hues, and colors applied with subtle realism rather than saturated modern grading. Keep all original detail intact.
Rescue a blurry or low-resolution photo
This photo is [PROBLEM: SLIGHTLY BLURRY / VERY LOW RESOLUTION / HEAVILY COMPRESSED] and it is the only shot of [SUBJECT]. Enhance it: recover sharpness and detail in the faces first, clean the compression artifacts, and produce a version that prints well at [SIZE]. Do not invent features that change what anyone looks like: stay faithful to the visible information.
Fix old flash and color cast damage
This [ERA: 1990S / 2000S] snapshot has period problems: [ISSUES: HARSH DIRECT FLASH FLATTENING THE FACES / RED-EYE / THE ORANGE CAST OF AGING PRINTS / A DATE STAMP IN THE CORNER]. Correct them all: natural balanced lighting on faces, accurate color, red-eye fixed, date stamp removed. Keep the authentic snapshot framing and feel, just technically healthy.
Composite the best version of a moment
These [NUMBER] photos are the same scene shot seconds apart, each with a flaw. Build the best single version using this photo as the base: take [ELEMENT: DAD’S SMILE] from the second photo and [ELEMENT: THE KIDS LOOKING AT CAMERA] from the third. Blend them seamlessly with consistent lighting and no visible seams. One photo where everyone looks their best.
Beyond correction lies transformation: changing seasons, styles, and moods while keeping the photo recognizably yours. These prompts push further while maintaining control.
Change the season or weather
Transform this photo from [CURRENT: SUMMER AFTERNOON] to [TARGET: FIRST SNOW OF WINTER / GOLDEN AUTUMN / SPRING BLOOM]. Change everything the season touches: foliage, ground cover, sky quality, light color, and how people’s clothing reads if visible. Keep the location, composition, and any people identical. The same place, a different month.
Relight the scene
Relight this photo as if shot at [NEW TIME: GOLDEN HOUR / BLUE HOUR / OVERCAST NOON / NIGHT WITH WARM INTERIOR LIGHT]. Recalculate the full lighting logic: sun direction and shadow length, color temperature, sky gradient, and how light falls on the subject. The scene and subject stay identical, only the light changes. It should feel like the photographer waited for better light.
Apply a photographic style
Regrade this photo in the style of [STYLE: KODAK PORTRA FILM / MOODY TEAL-ORANGE CINEMA / CLEAN BRIGHT EDITORIAL / FADED VINTAGE MATTE]. Adjust color grading, contrast curve, and grain character to match that style, but keep the content, sharpness, and detail intact. Show restraint: the grade should feel professional, not like a heavy filter.
Turn a photo into art
Reinterpret this photo as [MEDIUM: A WATERCOLOR PAINTING / A PENCIL SKETCH / A FLAT VECTOR ILLUSTRATION / AN OIL PAINTING]. Preserve the likeness and composition faithfully: anyone who knows the subject should recognize them instantly. Apply the medium’s authentic characteristics: [SPECIFICS: SOFT PIGMENT BLEEDS AND WHITE PAPER MARGINS / VISIBLE BRUSHWORK AND IMPASTO HIGHLIGHTS].
Build an edit recipe you can reuse
I edit many photos of the same type: [TYPE: PRODUCT SHOTS FOR MY STORE / FAMILY PHOTOS / LISTING PHOTOS]. Based on the edits we made to this one, write the reusable instruction paragraph I can paste with each new photo, covering: what always gets fixed ([COMMON FIXES]), the style target ([LOOK]), and what must never change ([PROTECTED ELEMENTS]). Then demonstrate it works by applying it to this next photo.
Photoshop requires you to execute edits manually with tools and layers; Nano Banana executes them from plain-language description. For object removal, background swaps, relighting, and restoration, describing the change is dramatically faster than doing it by hand. Photoshop still wins for pixel-exact production work and layered files, and many editors use Nano Banana for the heavy transformation, then finish in Photoshop.
The key technique is stating the preservation explicitly: name the change, then instruct that everything else stays exactly as it is. Nano Banana is notably strong at this kind of targeted edit, which is why it works for real photos of people and products where identity must survive the edit. If something drifts, a follow-up edit pointing at the drift usually restores it.
Yes, restoration is one of its most impressive use cases: repairing tears and stains, colorizing black and white prints, and recovering detail from faded or blurry originals. The honest caveat is that heavily damaged areas are reconstructed plausibly rather than recovered literally, so treat restored faces in destroyed regions as an informed approximation.
AI Prompts for Nano Banana Image Prompts: Photorealistic Portraits and AI Photography
Nano Banana (Google's Gemini image generation model, also known as Imagen 3) produces some of the most photorealistic portrait and product photography available from any AI image tool.
See promptsAI Prompts for AI LinkedIn Headshot Generation
A weak LinkedIn photo costs you interviews, deals, and credibility before you say a word.
See promptsAI Prompts for Instagram Profile Picture Generation
Your Instagram profile picture is the first thing every potential follower sees, appearing at 110 pixels in the grid and even smaller in Stories and search.
See prompts