← AI prompts · Poster

Mirror selfie with cartoon stickers — chibi AI prompt

A cute social poster — one real mirror-selfie surrounded by four chibi cartoon sticker versions of you, with doodles and handwritten words. The hard part is that all five figures have to be the same person, in two completely different rendering styles. Copy the free prompt below.

By Naresh Basude · Founder, Dressing School

AI mirror-selfie poster of a man in a cream shirt holding a phone, surrounded by four cute chibi cartoon sticker versions of himself with doodles — Made with the prompt below — a 4:5 vertical mirror-selfie sticker poster.

The one line that decides this: the uploaded photos are the only source of facial identity for both the real person and the surrounding cartoon versions — which is what stops the four stickers coming back as cute strangers standing around your photo.

Match the pose →

Create a highly detailed, photorealistic mirror-selfie poster portrait using the uploaded face-reference photographs as the only source of facial identity for both the real person and the surrounding cartoon versions. Replace the person in the composition with the individual from the uploaded identity references while preserving their exact facial structure, face shape, skin tone, forehead, hairline, hairstyle, eyebrows, eyes, nose, lips, ears, jawline, beard or moustache pattern, and any permanent visible features. The final result must clearly look like the person from the uploaded face-reference photos, not like the person shown in the reference image. Recreate the reference as a vertical 4:5 composition with a warm, cute, playful social-media poster aesthetic. The final image should combine one central realistic mirror-selfie portrait with multiple surrounding chibi / cartoon sticker versions of the same person. Keep the overall arrangement, mood, scale relationships, and visual balance closely aligned with the reference image. Place the main realistic subject in the center of the frame, slightly lower than mid-height, occupying most of the vertical space from roughly upper thigh or hip level up to the head. The subject should stand casually in front of a mirror, body turned slightly toward frame-left, with a relaxed posture and a slight lean. The head should angle slightly down and toward the phone, but most of the face should be obscured by the smartphone. The visible lower half of the face should preserve the user's identity, including jawline, cheek shape, skin tone, lips, and any visible facial hair. The subject's right arm should be raised in front of the face, holding a black smartphone vertically in a mirror-selfie position, covering the upper face almost completely. The fingers should wrap naturally around the phone, with a visible ring on one finger if shown in the reference. The left arm should hang naturally downward with the left hand tucked casually into the trouser pocket. The shoulders should remain relaxed and natural. Keep the stance easy, stylish, and spontaneous rather than formal. Dress the main subject in a loose, oversized, off-white to cream button-up shirt with long sleeves rolled or pushed casually up toward the forearms. The shirt should have a soft cotton or cotton-blend texture with natural creases, folds, and drape. Leave the top few buttons open for a relaxed styling effect, and keep the shirt slightly untucked or loosely hanging over the trousers. Pair it with dark charcoal or black trousers. Include the visible accessories from the reference: a dark wristwatch on the left wrist, thin bracelets on the right wrist, and a subtle chain necklace if visible near the collar opening. Do not add extra accessories beyond those visible. Preserve the subject's hairstyle from the uploaded identity references, but style it in the same general way as the reference image: short-to-medium length, dense, curly or wavy hair with volume, especially around the top and front, forming a youthful textured silhouette. If the identity references show glasses, preserve them in the cartoon versions and, where visible, in the realistic figure as appropriate. Maintain the user's true facial-hair pattern in both the realistic and cartoon versions. Surround the realistic central figure with four cute chibi cartoon sticker versions of the same person, each designed as a stylized mini character with a large head, small body, rounded features, expressive eyes, and a playful glossy 3D-cartoon finish. These cartoon figures must still clearly resemble the uploaded identity subject. Each one should have a clean white sticker-like outline around the body. Place them in approximately the same locations as the reference: 1. Top-left corner: a joyful chibi version floating or jumping with one arm raised in a waving gesture, smiling happily. 2. Top-right corner: a seated chibi version with crossed or bent legs, holding a drink cup with a straw, looking cute and relaxed. 3. Bottom-left corner: a seated chibi version with one eye winking, resting the face on one hand in a playful pose. 4. Bottom-right corner: a standing chibi version smiling and making a peace sign with one hand. All cartoon characters should wear a simplified version of the same outfit as the main subject: a cream shirt, dark trousers, and light shoes. Their hair, glasses if applicable, and facial cues should echo the real subject's identity. Keep them adorable, polished, expressive, and consistent in style. Use a mirror or reflective indoor background with warm beige, tan, and light brown tones, similar to an elevator interior, wardrobe mirror area, or softly lit indoor wall panels. The setting should remain simple and blurred enough not to distract, but still clearly suggest a mirror-selfie environment. Add a thin soft outline around the main realistic figure, similar to the sticker-outline effect in the reference, so the subject stands out from the background. Decorate the background with cute white hand-drawn doodles and handwritten text, placed around the main subject and cartoon stickers in the same general arrangement as the reference. Include sparkles, stars, tiny hearts, a small camera doodle, a drink doodle, and playful curved accent lines. Include the handwritten-style words: * "shine"* near the upper-left * "bright day" near the top-center * "happy" on the left side * "smile♡" on the right side Keep the text in a white hand-drawn script style, playful and youthful, integrated into the composition rather than looking like printed typography. Preserve the cute scrapbook / sticker-poster feeling. Use a warm indoor lighting setup with soft, flattering illumination from above and slightly in front, producing gentle highlights on the shirt, hands, hair, and exposed skin. Keep the lighting warm beige-golden, with soft shadows and no harsh contrast. The mirror reflection look should feel natural and believable. Use subtle glow and smooth tonal transitions to reinforce the cozy, cheerful mood. The overall colour palette should stay within warm beige, cream, caramel, light brown, taupe, soft black, and muted gold, with the white doodles and outlines providing contrast. Apply a cozy, slightly edited social-media colour grade with soft warmth, moderate contrast, clear details, and a polished but natural finish. Use a realistic portrait camera perspective for the central figure, roughly eye-level to slightly chest-level in the mirror, with a medium shot framing. The image should feel as though it was captured casually on a phone in a mirror, while the added cartoon sticker elements feel intentionally composited into a cute editorial design. Keep the central realistic figure sharp and detailed, while the background remains slightly softer. Keep all doodles, outlines, and cartoon elements crisp and readable. The overall mood should be cute, youthful, playful, cheerful, cozy, and stylish, blending a real mirror-selfie photo with adorable chibi character stickers of the same person. Maintain strong identity fidelity throughout. The final result should look like a premium, high-resolution, professionally designed social-media poster based on the uploaded identity photos and the provided reference layout. NEGATIVE PROMPT Wrong person, copied face from the reference image, identity loss, altered face shape, changed skin tone, incorrect hairstyle, incorrect hairline, incorrect beard or moustache pattern, wrong glasses, missing curly or wavy hair, generic face, distorted jawline, exaggerated facial features, bad anatomy, extra fingers, extra hands, deformed hands, broken wrist, unnatural arm positions, incorrect selfie pose, phone not covering face, face fully visible when it should be partially covered, wrong hand in pocket, wrong clothing, tight shirt, wrong shirt color, patterned shirt, extra accessories, missing watch, missing bracelet, missing ring, formal suit, jacket, shorts, bright colored clothing, wrong trousers, outdoor background, street background, bedroom clutter, complex scenery, dark background, neon lighting, cold lighting, harsh shadows, dramatic cinematic shadows, overexposed highlights, low resolution, blur, noise, pixelation, artifacting, messy composition, missing cartoon stickers, too many cartoon stickers, inconsistent cartoon style, cartoon stickers not matching the subject identity, no white outlines, missing doodles, missing handwritten words, wrong text, unreadable text, extra text, watermark, logo, signature, border, frame, poster brand, exaggerated 3D rendering, anime style, painterly style, flat illustration, scary expression, sad mood, aggressive pose, distorted mirror effect, duplicate limbs, misaligned features, poor typography, cluttered layout, or low-quality social-media edit.

About this prompt

Almost every other prompt on this site asks a model to do one thing well. This one asks for two incompatible things in one frame: a photographic mirror selfie of you, and four glossy 3D cartoon miniatures that also have to be you. The cream shirt, the flat light and the white outlines all exist to let those two layers share a picture.

The line doing the most work is the first one: the uploaded photos are the only source of facial identity for both the real person and the surrounding cartoon versions. Without it you are running two unrelated jobs — an identity-preserving portrait, and four generic cute characters — with nothing tying them together.

The lighting is deliberately boring: warm indoor light from above and in front, soft shadows, no harsh contrast. A chibi sticker is flatly lit by nature, and a hard-lit photoreal figure beside four flat cartoons reads as two images stuck together. Thin white doodles need that even range too — they have nowhere to sit in a blown highlight or a black shadow.

The wardrobe is a value decision, not a fashion one. The shirt is off-white to cream, the room is beige and tan, and the palette paragraph says the white doodles and outlines provide the contrast. Everything sits one notch below white so pure white is left for the line work. A true white shirt spends the brightest value in the picture on the very surface the figure's own outline has to separate from.

The composition is built around the corners. The subject sits slightly lower than mid-height so the top two stickers get headroom, and each is enumerated separately, corner and pose named, because AI image tools have a well-known habit of losing count when asked for a set number of repeated objects. Other clauses point at "the reference" — a ring if shown in the reference, doodles in the same general arrangement as the reference — because this was written to sit beside a layout image. With face photos alone the layout rests entirely on the explicit lines: 4:5, centred figure, four named corners, the doodle word list.

How to use it

Tap Copy prompt above.
Open a tool that takes reference photos — Gemini, Nano Banana, ChatGPT in image mode, Midjourney with a character reference, Leonardo.
Attach two to four sharp photos: one showing your jaw, mouth and facial hair clearly, one front-on shot showing your hairline and hair volume.
Add the layout reference image too if you have one, then generate at 4:5.

Why these lines are in the prompt

most of the face should be obscured by the smartphone — with the visible lower half of the face should preserve the user's identity, this halves the identity problem. Eyes and nose are where a face stops being yours; the phone removes them and leaves recognition to the jaw, mouth and facial hair.

the left hand tucked casually into the trouser pocket — one hand deleted from the drawing job. The other is already wrapped around a phone in front of a face, the most demanding anatomy in the picture.

a clean white sticker-like outline around the body, and then add a thin soft outline around the main realistic figure — the same device on both layers, so the photograph is treated as a sticker too and the poster reads as one designed object rather than cartoons dropped onto a photo.

all cartoon characters should wear a simplified version of the same outfit as the main subject — the outfit is the second identity signal, and the one that survives simplification. When a chibi head cannot carry recognition alone, five figures in the same cream shirt still read as one person.

simple and blurred enough not to distract — the background carries four stickers, a scatter of doodles and four handwritten words on top of it. Put texture underneath and the white line work stops being readable.

The three ways this goes wrong

The phone slides off your face. You get a fully visible face with a phone floating near it. The prompt's negative list names this twice — phone not covering face and face fully visible when it should be partially covered — a fair sign of how fragile it is. Negatives are weak medicine here, so state the boundary positively: "The phone must cover the face from the eyebrow line down to just above the upper lip. Only the mouth, jaw and chin remain visible."

The handwritten words come out wrong. Four separate strings, one pairing a word with a heart glyph, and AI image tools have a well-known habit of mangling text. Ask for less: "Write one handwritten white word, shine, in the upper-left area. Do not place any other text anywhere in the image." If a stray asterisk turns up rendered into the poster, that is the list punctuation being picked up as text — rewrite those lines as a plain sentence.

The four stickers are cute strangers. Adorable, consistent with each other, and not you — and the poster only means anything if the cartoons are recognisable. After the numbered corner list, add: "Each chibi must keep the same hairline shape, hair volume and texture, eyebrow shape, skin tone and facial-hair pattern as the uploaded photos. Simplify the features, do not replace them."

Making it yours

Your name in the handwriting. Replace the four doodle words with one: "Write one handwritten white word, my name Rohit, near the upper-left." A single short string is the version most likely to come back spelled right.
A different wardrobe. Swap the shirt paragraph for what you actually wear — "a charcoal oversized t-shirt with light-wash jeans". Going dark, add "keep the white outline thin"; going light, stay off-white so the outline still separates.
A different mirror. A gym mirror, a lift lobby, a tailor's fitting room. Keep both conditions — simple, and blurred — and keep the warm tones the doodles depend on.
Two stickers instead of four. Cut the numbered list to two on opposite corners, top-left and bottom-right. Diagonal placement holds the frame together; two on one side tips it.
4:5 to 9:16 for a story. It works, but the corners pull away from the figure and the stickers look like they are drifting rather than orbiting it. Re-specify them as a pair above the head and a pair beside the hips.
Uncover the face. Delete the phone-obscuring lines and hold the phone at chest height. You get a full face, and take back the hardest part of the job: your real face now has to be right as well as four cartoon versions of it.

If the typography half interests you more, the Ganpati invitation card prompt renders words inside the image too — a date, a time and a venue set at display size rather than handwritten.

Then match the pose with your camera using the Pose Mirror, sharpen your beard styling, get the shoes for the look, or browse more AI photo prompts.

Frequently asked questions.

How many photos should I upload, and what should be in them?

Two to four, picked for this image rather than for a normal portrait. The phone hides your upper face in the realistic figure, so at least one photo needs a clear view of your jaw, mouth, chin and facial hair. The chibi versions lean harder on hair, so include one front-on shot showing your hairline and hair volume.

Do I need to set identity strength or transformation strength?

ChatGPT, Google Gemini and Nano Banana do not expose transformation-strength, identity-strength or beautify sliders at all, so in those three there is nothing to set. Say it in words instead: ask the tool to keep your facial structure exactly as it is in the uploaded photos, and not to smooth, slim or beautify the face.

I wear glasses — will they survive the cartoon versions?

The prompt handles glasses conditionally: if the identity references show them, they carry into the cartoon versions. They survive stylisation better than most features, because a chibi head keeps very few landmarks and a frame is one of them. The real problem is inconsistent references — if half your uploads have glasses, expect them on some figures and not others.

Can I use this as a profile picture?

Not as it stands. Profile pictures get cropped to a circle or a square, and the four stickers sit in the four corners, so that crop removes all of them. For the avatar slot, generate a single chibi instead: keep the identity paragraph, delete the mirror-selfie and doodle sections, and ask for one chibi on a plain warm beige background.

Why do the watch and bracelets end up on the wrong wrists?

Because the prompt names them by your left and right, and this is a reflection. AI image tools have a well-known habit of swapping sides when a pose is described from the subject's point of view rather than from the frame. Tie each accessory to an arm by its job instead: the bracelets on the raised arm holding the phone, the watch on the wrist of the hand in the pocket.

Is this prompt free, and what do I need to run it?

Yes, it is free to copy and use. You need an AI image tool that accepts reference photos and two to four clear pictures of your own face. Budget a few attempts, though: five versions of one person plus hand-drawn text in one image is far more to get right at once than a plain portrait.

Pair it with

Get the mirror-selfie pose right on camera.

Got the look? Match the pose with your camera.

Open Pose Mirror →