15 August Photo Prompts: AI Prompts That Look Real

A good 15 August photo prompt is the difference between a portrait that looks staged and one that looks like it was actually shot that morning. Most people paste one line into Gemini or ChatGPT, get a flag that looks painted on, and give up.

The usual problem isn’t the AI tool. It’s the prompt. Vague instructions like “person with Indian flag, patriotic photo” leave the model to guess the flag’s colors, the lighting, and the pose, and it guesses wrong more often than you’d expect.

This guide has seven tested prompt versions: a base version, an AI-general version, a ChatGPT-tuned version, a viral-style version, a Gemini version, plus a boy version and a girl version. Every prompt below went through multiple rounds of edits before it made the cut.

I tested every prompt in this guide on Gemini and ChatGPT, most of them more than 5 times, before writing any of it down. This guide is for beginners, content creators, and anyone who wants their Independence Day photo to look real instead of painted on. If it saves you a few failed generations, send it to a friend or family member who’s about to try the same thing.

15 August Photo Prompt

I started with the shortest version I could write: just “person holding the Indian flag, 15 August.” Gemini gave me a flag with six stripes and stars that don’t exist on it. ChatGPT did better but placed the flag upside down twice in a row. The fix was naming the flag’s exact band order and the Ashoka Chakra’s 24 spokes directly in the prompt, not leaving it implied.

Use the uploaded photo as the only reference for facial identity. Keep the face, skin tone, hairstyle, and expression exactly as uploaded. Do not smooth skin, change age, or alter facial features.



Create a photorealistic vertical portrait (4:5) of the same person standing on a sunlit rooftop terrace on the morning of 15 August. The person wears a plain white kurta with a small tricolor pin on the chest, positioned above the heart. They hold a full-size Indian national flag on a wooden pole with their right hand, the flag gently moving in a light breeze: saffron band on top, white band with a navy blue Ashoka Chakra of 24 spokes in the center, green band at the bottom. The flag fabric shows natural folds and slight motion blur only on the loose end.



Background: a soft-focus skyline with a few other flags visible on nearby buildings, warm morning light coming from the left at a low angle, light haze, no crowd, no other people.



Camera: 50mm lens, f/2.2, natural light, eye-level angle, shallow depth of field, sharp focus on the face and flag, realistic skin texture, no beauty filter, no illustration style, no watermark.

Customization tips:

• Swap the kurta for a saree or a plain shirt by changing only that one line.
• If the flag looks flat, add “flag catching wind, fabric slightly curved” right after the color description.
• For evening shots, change “warm morning light” to “golden hour light, sun low on the horizon.”
• Add “no text overlay” at the end if the AI keeps writing random captions on the image.

15 August AI Photo Prompt

Gemini handled group photos worse than solo ones in my tests. Ask for three people holding one flag together and at least one face gets distorted almost every time. Splitting the instruction into a lead subject plus two supporting figures cut that failure rate by a lot.

Use the uploaded photo as the only facial identity reference for the main subject in the center. Keep their face, skin tone, and expression unchanged.



Create a photorealistic horizontal photo (3:2) of a family of three standing together outdoors on the morning of 15 August: the main subject in the center, one person to their left, one to their right. All three look toward the camera with relaxed, natural smiles, not posed stiffly. The center subject holds a medium Indian flag upright in front of the group at chest height, with the saffron, white (24-spoke navy Ashoka Chakra), and green bands correctly ordered top to bottom. The two side subjects each rest one hand lightly on the flag’s pole for support.



Clothing: simple everyday outfits in white, off-white, or light beige, no clashing prints, one small tricolor badge visible on the center subject only.



Background: a quiet park with a flagpole in soft focus behind the group, green grass, morning light from the right, mild lens flare.



Camera: 35mm lens, f/2.8, natural daylight, eye-level, medium depth of field so all three faces stay sharp, documentary-style realism, no cartoon effect, no extra fingers, no duplicate people.

Customization tips:

• For two people instead of three, delete the “one person to their left” line and change “family of three” to “two people.”
• If faces blur at the edges, ask for a smaller aperture, meaning “f/4” instead of “f/2.8,” to keep more of the frame sharp.
• Change “quiet park” to “school courtyard” or “home balcony” depending on where the real photo will be used.
• Keep the clothing instructions short. Long clothing paragraphs are where AI tends to invent random accessories.

15 August ChatGPT Photo Prompt

During testing, ChatGPT occasionally changed the person’s facial features even when I clearly asked it to preserve the uploaded face. The most reliable solution was to explicitly state that the uploaded image is the only facial identity reference, then list the facial features that must remain unchanged. Placing “no text, no caption, no watermark” at the very end of the prompt also prevented unwanted text in almost every test.

caption, no watermark” at the very end of the prompt fixed it almost every time after that.

Use the uploaded image as the ONLY facial identity reference. Match the person’s face exactly, including facial structure, eye shape, eyebrows, nose, lips, jawline, skin tone, hairstyle, hair volume, hairline, makeup level, and apparent age. Do not beautify, slim the face, enlarge the eyes, change the hairstyle, or alter any facial proportions. Preserve realistic skin texture and natural facial details.



Create a photorealistic vertical portrait (4:5) of the same person standing outdoors at dusk on 15 August. Keep the camera at eye level, with the person occupying approximately the same frame size as a portrait photograph. The face should be looking directly at the camera with a natural closed-mouth smile.



The person wears the same white outfit as in the reference photo. Long black wavy hair remains unchanged.



The person holds a small handheld Indian national flag in one raised hand, elbow slightly bent. The flag must be correctly rendered with saffron on top, white in the middle with a navy-blue 24-spoke Ashoka Chakra, and green at the bottom. The flag should be proportionally small and realistic, not oversized.



Background: warm dusk atmosphere with soft string lights and a generic softly blurred illuminated domed building silhouette (not any identifiable monument). Blue-orange evening sky with a few faint stars. Background remains heavily blurred (bokeh).



Camera settings: 50mm lens, f/1.8, shallow depth of field, sharp focus on the eyes and face, realistic lighting, natural skin texture, cinematic color grading, high dynamic range.



Negative prompt: do not change the face, do not generate a different person, no beauty filter, no face swap artifacts, no AI-looking skin, no oversized flag, no distorted hands, no extra fingers, no extra limbs, no duplicate flag, no text, no watermark, no illustration, no painting, no cartoon style.

Customization tips:

• If the face doesn’t match the uploaded photo, start the prompt with “Use the uploaded image as the ONLY facial identity reference. Match the face exactly.”
• For a more natural result, add “shot on a phone camera, realistic DSLR photo, unedited skin texture” near the beginning of the prompt.
• Replace “standing outdoors at dusk” with “standing in a lush green garden”, “walking on a rooftop terrace”, or “standing beside a balcony railing” to create different Independence Day scenes while keeping the same facial identity.
• Keep the negative prompt focused on identity preservation, such as “do not change the face, no beauty filter, no illustration, no text, no watermark.” Too many negative instructions can reduce consistency instead of improving it.

15 August Kid Patriot AI Photo Prompt

When testing this prompt, I noticed that AI models often changed the child’s face, made the flag unrealistically large, or placed the camera at a normal eye level instead of the dramatic low-angle view. The most consistent results came from describing the camera angle, subject size, and flag proportions in detail while asking for a natural park background with soft bokeh. Keeping the instructions focused on realism also produced much better facial expressions and natural lighting.

Use the uploaded image as the ONLY facial identity reference. Preserve the child’s facial structure, skin tone, hairstyle, expression, apparent age, and all facial proportions exactly as shown. Do not beautify, mature, or alter the child’s identity.



Create a photorealistic vertical portrait (4:5) of the same little girl standing proudly in a beautiful green park on Independence Day.



Capture the image from a low-angle (camera positioned near the ground, looking upward) to create a heroic perspective while keeping the girl’s face as the main focus.



The child wears the same white outfit as the reference image and smiles naturally while looking directly into the camera.



She proudly holds a realistic Indian national flag on a white pole using both hands. The flag should have saffron on top, white in the middle with a navy-blue 24-spoke Ashoka Chakra, and green at the bottom. The flag should appear naturally sized with gentle movement in the breeze.



Background should feature a lush green park with tall trees, soft sunlight filtering through the leaves, blurred walking paths, and creamy bokeh. No crowds or distracting objects. The environment should feel peaceful, bright, and natural.



Lighting should be warm natural daylight with soft highlights on the face. Use 50mm DSLR photography, f/1.8 aperture, shallow depth of field, HDR, ultra-realistic skin texture, natural colors, high detail, professional portrait photography.



Negative Prompt: no illustration, no cartoon, no painting, no beauty filter, no face swap artifacts, no AI-looking skin, no oversized flag, no distorted hands, no extra fingers, no duplicate limbs, no blurry face, no text, no logo, no watermark.

Customization Tips

• Add “golden hour sunlight” for a warmer and more cinematic Independence Day look.
• Replace “green park” with “flower garden”, “tree-lined avenue”, or “nature trail” for different outdoor scenes while keeping the same composition.
• If the face changes, move “Use the uploaded image as the ONLY facial identity reference” to the very first line of the prompt.
• If the camera angle looks flat, specify “camera placed close to the ground, looking upward from below the child’s eye level.”
• To create a stronger portrait effect, add “the face occupies approximately one-third of the frame with sharp focus on the eyes.”
• Always finish the prompt with “no text, no caption, no watermark” to avoid unwanted overlays in the final image.

15 August Viral Photo Prompt

The versions that actually got shared more in my small test group weren’t the most detailed ones. A slightly candid, mid-laugh photo outperformed the perfectly posed version in every side-by-side comparison I showed people.

Use the uploaded photo as the only facial identity reference. Preserve the exact face, skin tone, and features.



Create a realistic vertical photo (4:5) of the person mid-laugh, head tilted slightly back, eyes crinkled in genuine amusement, standing outdoors on 15 August with a light breeze. Tricolor paper confetti (saffron, white, green) is caught mid-fall around them, some pieces blurred from motion. They hold one small paper flag loosely in one hand, arm relaxed at their side rather than raised.



Clothing: casual white t-shirt with a small tricolor wristband, nothing else patriotic-themed, to keep the focus on the expression.



Background: a bright, slightly overexposed sky, soft bokeh of a distant crowd, no visible faces in that crowd.



Camera: 85mm lens, f/1.8, bright natural daylight, candid street-photography style, slight motion blur on the confetti only, sharp focus on the face, no posed stance, no forced smile, no text overlay.

Customization tips:

• Motion blur on the confetti only, not the face, is the detail that sells the candid look. Keep that line even if you cut others.
• For a rooftop version, change “distant crowd” to “city skyline far in the distance.”
• If the laugh looks fake, add “natural, unposed, caught off guard” right before the expression description.
• Generate this one at least twice. The first draft often reads posed even when the prompt says candid.

Gemini Independence Day Photo Prompt

Gemini follows layered background instructions better than the other tools in my tests, so this version leans on that strength with a busier, more detailed scene than I’d try in ChatGPT.

Use the uploaded photo as the only reference for facial identity. Keep the face, skin tone, hairstyle, and natural expression exactly as uploaded, no beautification.



Create a photorealistic vertical image (4:5) of the person standing at a small flag hoisting ceremony on the morning of 15 August. They stand slightly to the right of a tall flagpole, looking up and to the left at a large Indian flag mid-hoist, caught in a light breeze, correct band order of saffron, white with a 24-spoke navy Ashoka Chakra, and green. Their right hand rests near the rope, not pulling it, since the flag is already mostly raised.



Around them, in soft focus: a handful of other people looking up in the same direction, none facing the camera, a couple of folding chairs, a small table with marigold garlands.



Lighting: early morning sun low on the horizon, warm golden tones, soft long shadows across the ground, light haze typical of an August morning.



Camera: 35mm lens, f/2.5, eye-level, medium depth of field, sharp focus on the main subject’s face, everyone else softly blurred, documentary realism, no distorted flag, no incorrect chakra spokes, no crowd facing camera, no text overlay.

Customization tips:

• Gemini responds well to naming exact objects in the background, like “folding chairs” and “marigold garlands,” instead of vague words like “decorations.”
• If the flag looks too small, add “flag fills the upper third of the frame” as its own sentence.
• Swap “flag hoisting ceremony” for “school assembly” for a younger-feeling setting.
• Keep the Ashoka Chakra spoke count in the prompt every time. Leaving it out is the single biggest reason the flag comes out wrong in Gemini.

15 August Boy Photo Prompt

Boys’ outfits kept losing detail in my early drafts. The jacket would turn into a plain shirt by the second generation. Naming the fabric and one specific stitching detail fixed it.

Use the uploaded photo as the only facial identity reference for the boy. Keep his exact face, skin tone, hairstyle, and age. Do not age him up or change his features.



Create a photorealistic vertical portrait (4:5) of the boy standing outdoors on the morning of 15 August, wearing a plain white cotton shirt with the sleeves rolled to the elbow and a thin tricolor thread bracelet on one wrist. He holds a small paper or cloth flag in one hand, resting against his shoulder rather than raised high. His expression is a relaxed, natural smile, looking directly at the camera.



Background: a school building entrance softly blurred, a few balloons in saffron, white, and green tied loosely to a railing, morning light from the left.



Camera: 50mm lens, f/2, natural daylight, eye-level, shallow depth of field, sharp focus on the face and shirt texture, realistic skin texture, no beauty filter, no oversized head, no extra fingers, no text overlay.

Customization tips:

• The rolled-sleeve detail keeps the look casual. Remove it for a more formal school-uniform version instead.
• Swap the balloons for a hand-painted poster in the background for an older-looking teen version.
• If the flag looks too stiff, add “flag slightly wrinkled from being carried” for a more lived-in feel.
• Keep the age unchanged in your instructions. AI models tend to age young faces up slightly unless you say otherwise.

15 August Girl Photo Prompt

Girls’ photos in my tests came out over-smoothed by default, skin looked plastic even when I never asked for that. Adding one line about visible skin texture fixed almost every generation after that.

Use the uploaded photo as the only facial identity reference for the girl. Keep her exact face, skin tone, hairstyle, and natural features. Do not smooth her skin, slim her face, or apply any beauty filter.



Create a photorealistic vertical portrait (4:5) of the girl standing outdoors on the morning of 15 August, wearing a simple white kurta with a small tricolor dupatta draped loosely over one shoulder. She holds a small flag close to her chest with both hands, looking slightly off to the side with a calm, thoughtful expression rather than a big smile.



Background: a garden with a few marigold flowers in soft focus, warm afternoon light, gentle shadows.



Camera: 85mm lens, f/1.8, natural daylight, eye-level, shallow depth of field, visible natural skin texture including minor texture and pores, sharp focus on the face, no plastic skin, no over-smoothing, no oversized eyes, no beauty filter, no text overlay.

Customization tips:

• The “visible natural skin texture” line matters more here than in most other prompts. Most default outputs over-smooth young female faces.
• Swap the calm expression for “mid-laugh, eyes crinkled” for a more candid version, similar to the viral prompt above.
• Change the dupatta color detail to match the real reference photo, and keep everything else the same.
• Garden backgrounds work better than street backgrounds for this one in my tests, with fewer distorted background faces.

Flag mistakes AI models keep making

Across every tool I tested, the same handful of flag errors kept showing up: the wrong number of spokes on the Ashoka Chakra, the color bands in the wrong order, stray stars or symbols that don’t belong on the flag, and flags that look stiff and printed instead of made of fabric.

The fix is the same each time. State the band order explicitly (saffron, white, green, top to bottom), name the chakra’s 24 spokes, and add one line about fabric movement, like “flag catching a light breeze.” Vague words like “Indian flag” alone leave too much to chance, even on a second or third generation.

Right aspect ratio for where you’re posting

The prompts above default to 4:5 for a reason: it’s the safest ratio for an Instagram feed post and still crops cleanly for most other platforms.

• Instagram or Facebook feed post: 4:5 vertical.
• Instagram or WhatsApp Status story: 9:16 vertical, taller than the feed version.
• WhatsApp DP or profile picture: 1:1 square, so keep the subject centered with some space around them.
• Facebook cover or a wide banner: 3:2 or 16:9 horizontal, which is why the AI photo prompt version above uses a horizontal layout.

Change the ratio inside the prompt itself, not after generating. Cropping a 4:5 image down to a square after the fact often cuts off the flag or a hand.

Conclusion

The biggest lesson from testing dozens of 15 August photo prompt variations is that specific beats short every time. A vague prompt leaves the AI guessing at the flag, the lighting, and the pose, and it guesses wrong more often than not.

Start with the version closest to your photo, whether that’s the boy, girl, viral, or ceremony style. Change one detail at a time between generations rather than rewriting the whole prompt. Try the same prompt across Gemini and ChatGPT too, since they don’t fail in the same places.

One takeaway to use right now: whichever prompt you copy, name the flag’s exact band order and the 24-spoke chakra yourself. That single line fixes more bad generations than anything else in this guide.

Also Read:

10 Beautiful ChatGPT Saree Photo Editing Prompts for Girls

Instagram Trending Funny Baggy Jeans AI Photo Editing Prompts