Mera prompt simply means “my prompt” in Hindi. Below are copy-ready style prompts for the looks people share on Instagram and Reels, plus an extractor for the case where you saved the photo but never got the text.
Last updated: November 2026. Styles are reviewed when the surrounding trend changes.
Drop the saved image here
or click to choose a file
JPG, PNG, WEBP up to 4MB
The extractor describes what it can see. It cannot know the creator's private prompt and is not an exact reproduction of their result.
Upload a photo and the extracted prompt will appear here. You can then copy it in the format your AI tool expects.
Each card is a prompt template. Replace the example details, keep the face-lock sentence first, and change one thing at a time.
Clean head-and-shoulders shot for a profile picture.
Use the uploaded photo as the face reference and keep the face, skin tone, and hairstyle exactly the same. Create a clean studio portrait of me in a smart casual outfit, standing in front of a soft grey gradient backdrop. Use a large softbox key light with gentle fill, a relaxed head-and-shoulders pose, and a calm natural smile. Crisp focus on the eyes, realistic skin texture, subtle editorial finish, square crop, no text, no logos.
Match-day poster with your name and number.
Use the uploaded photo as the face reference and keep my identity, face shape, skin tone, and hairstyle unchanged. Show me as a cricket fan in a packed night stadium, wearing a plain team-coloured jersey with your own name and your favourite number. Hold a bat over one shoulder, look toward the field, and place floodlights, flags, and a blurred cheering crowd behind me. Realistic stadium lighting, slight motion in the crowd, vertical sports-poster composition, no real team badges or sponsor logos.
Soft evening light with a saree portrait.
Use the uploaded photo as the face reference and keep my facial features, skin tone, and hairstyle exactly the same. Dress me in a a saree in a colour you choose with a fine zari border and simple gold earrings, hair tied back with jasmine. Place me on a rooftop at golden hour with a warm sky, soft rim light on the fabric, and a shallow-depth background. Three-quarter pose, calm expression, realistic fabric folds, cinematic colour grade, no text, no watermark.
Palace-style ethnic portrait for wedding season.
Use the uploaded photo as the face reference and preserve my identity, skin tone, and hairstyle. Create a royal Indian wedding portrait in an an embroidered lehenga or sherwani with layered jewellery, standing in a palace courtyard with marble pillars and marigold garlands. Warm evening light, soft shadows, regal posture, detailed fabric texture, shallow depth of field, ultra-realistic finish, no text, no logos.
Warm festival light with diyas and rangoli.
Use the uploaded photo as the face reference and keep my face, skin tone, and hairstyle unchanged. Show me in festive ethnic wear holding a lit diya on Diwali night, with rangoli on the floor, string lights, and rows of glowing diyas around me. Warm golden bokeh, soft key light on the face, a gentle smile, realistic skin texture, vertical portrait, no text, no logos.
Daylight action shot with colour powder.
Use the uploaded photo as the face reference and keep my facial features, skin tone, and hairstyle exactly the same. Show me celebrating Holi in bright daylight, wearing a white kurta splashed with pink, yellow, and green powder. Add a cloud of colour bursting around me, a natural laugh, slight motion in the fabric, and a bright outdoor background. Vivid colour, sharp detail, realistic skin texture, vertical framing, no text, no watermark.
Attitude shot on an empty highway.
Use the uploaded photo as the face reference and keep my identity, face, skin tone, and hairstyle unchanged. Show me in a black leather jacket and aviators beside a a matte motorcycle or car on an empty desert highway at sunset. Low camera angle, warm orange light on the dunes, long shadows, dust in the air, a confident stance. Photorealistic, cinematic colour grade, no brand logos, no text.
Two-photo edit for anniversaries.
Use the two uploaded photos as face references and keep both faces exactly the same, including features, skin tone, and hairstyles. Create a romantic portrait of us standing close together among deep red roses and candlelight, wearing elegant evening outfits. Warm rim light, rose petals in the air, soft background blur, natural expressions, ultra-realistic detail, vertical framing, no text, no logos.
Turn a portrait into a desk figurine.
Use the uploaded photo as the face reference and keep the face, hairstyle, and outfit recognisable. Turn me into a realistic 1/7 scale collectible figurine standing on a clear acrylic base on a wooden desk. Place a toy-style box printed with the same character beside the figure, and show a monitor with a simple 3D model behind it. Soft indoor light, product-photo realism, clean desk, no brand names, no watermark.
Vintage hand-drawn ad style with a prop.
Use the uploaded photo as the face reference and keep my identity recognisable while turning it into a hand-drawn caricature with the same hairstyle shape and skin tone. Show me smiling and holding a product toward the viewer, in a vintage Indian advertisement style with bold outlines and a warm mustard, teal, and cream palette. Add a large handwritten headline reading your own short line and a smaller tagline on the side. Clean cream background, centred composition, portrait poster format, no extra text, no watermark.
| What you need | A typical prompt list | This page |
|---|---|---|
| Start from a saved image | Not supported — you still have to guess the wording | Upload the image and extract a prompt you can edit |
| Keep your own face | Often left out, or buried in one line | Every template starts with a face-lock sentence |
| Know which tool to use | Usually one generic instruction | Switch between ChatGPT/Gemini, Midjourney, Flux, and Stable Diffusion |
| Understand the limits | Rarely explained | Failure cases and a clear “not the original prompt” note |
The phrase is simple, but the search hides a specific workflow: see a photo, get the text, reuse it with your own face.
In Hindi, “mera” means “my”. A mera prompt is the text someone used to make an image. The phrase is shorter than asking for “the prompt you used”.
On Instagram and Reels, viewers ask for the prompt under an AI photo. Creators reply with the text, and the next viewer searches for it. The search grew out of that repeated exchange.
A useful prompt is not one sentence. It describes the subject, outfit, pose, camera angle, lighting, background, finish, and the face-lock instruction. Remove one part and the result drifts.
Styles rotate with festivals, films, and cricket seasons. A Diwali portrait prompt will not produce a stadium photo, so treat a card as a starting point.
The workflow is the same whether you start from a copied style or an extracted prompt. The order matters more than the tool you choose.
Choose a card above and copy the full text. Keep the bracketed placeholders until you know what to change; a half-edited prompt is harder to debug.
Open Gemini or ChatGPT and attach a sharp, front-facing photo with even light. A re-saved screenshot, a heavy filter, or a turned-away face gives the model less to preserve.
Paste the prompt and send it. If your identity changes, repeat the instruction that the face, skin tone, and hairstyle must stay the same, then regenerate.
Adjust the outfit, background, or crop in separate runs. Rewriting everything after each result hides which change helped.
This is the gap the extractor fills. It reads the image, not the creator’s private notes.
The analysis describes the visible subject, style, lighting, colour palette, composition, details, and estimated aspect ratio. Those details become an editable prompt.
It cannot recover the exact text the creator typed, the model, the seed, the reference images, or any private instruction. An extracted prompt is a reconstruction, not a receipt.
Use it when you have a saved image and want a practical starting point, or when a style card is close but not close enough. Extract first, then edit the parts that do not match your photo.
Most failures come from the input photo or an overloaded prompt, not from a missing magic word.
The model may blend your face with the reference style. Repeat the identity instruction, use a sharper source photo, and change less of the scene per run.
Objects held toward the camera are hard. Describe the grip and the size, and expect to regenerate a few times before the hand looks natural.
Short names and numbers are easier than long lines. Put the exact text in quotation marks and check the spelling before you post.
Too many instructions produce a muddy image. Keep the subject, outfit, light, and background; move the rest to a second run.
Short answers to the questions people ask before copying a prompt.
Need a broader description instead of a prompt? Use AI Describe Image for scene details and OCR.
Start with a prompt card, or upload the photo you saved and extract a reusable prompt you can edit for your own face and setting.