You in a silk saree / kurta arranging diyas
Diwali AI photo prompt · 5 styles · works with Nano Banana, ChatGPT, Gemini, Midjourney & Grok
Scene: The person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk.
Style 1: Ultra-realistic photography
Ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field
Nano Banana prompt
Using the uploaded photo as reference, keep every face, facial feature and identity 100% identical. Create the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field. Highly detailed, 4K, aspect ratio 4:5.
ChatGPT prompt
Create an image using my uploaded photo. Show the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Style: Ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field. Keep the face exactly recognisable — do not change facial features, skin tone or hairline. Portrait (2:3) orientation.
Gemini prompt
Edit my photo into a new image: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field. Keep my face the same as in the original photo and make the result look natural, sharp and high quality.
Midjourney prompt
the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk, ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field, highly detailed --ar 4:5 --oref <your-photo-URL> --ow 100
Grok prompt
Edit this photo: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Ultra-realistic photography, warm golden-hour tones, 85mm lens, shallow depth of field. Keep my face exactly the same as in the photo. Aspect ratio 4:5, high detail.
Style 2: Cinematic movie-poster look
Cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light
Nano Banana prompt
Using the uploaded photo as reference, keep every face, facial feature and identity 100% identical. Create the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light. Highly detailed, 4K, aspect ratio 4:5.
ChatGPT prompt
Create an image using my uploaded photo. Show the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Style: Cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light. Keep the face exactly recognisable — do not change facial features, skin tone or hairline. Portrait (2:3) orientation.
Gemini prompt
Edit my photo into a new image: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light. Keep my face the same as in the original photo and make the result look natural, sharp and high quality.
Midjourney prompt
the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk, cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light, highly detailed --ar 4:5 --oref <your-photo-URL> --ow 100
Grok prompt
Edit this photo: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Cinematic movie-poster look, rich orange and magenta colour grading, dramatic rim light. Keep my face exactly the same as in the photo. Aspect ratio 4:5, high detail.
Style 3: Vintage 90s film photo
Vintage 90s film photo, soft grain, slightly faded colours, nostalgic Bollywood vibe
Nano Banana prompt
Using the uploaded photo as reference, keep every face, facial feature and identity 100% identical. Create the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Vintage 90s film photo, soft grain, slightly faded colours, nostalgic Bollywood vibe. Highly detailed, 4K, aspect ratio 4:5.
ChatGPT prompt
Create an image using my uploaded photo. Show the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Style: Vintage 90s film photo, soft grain, slightly faded colours, nostalgic Bollywood vibe. Keep the face exactly recognisable — do not change facial features, skin tone or hairline. Portrait (2:3) orientation.
Gemini prompt
Edit my photo into a new image: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Vintage 90s film photo, soft grain, slightly faded colours, nostalgic Bollywood vibe. Keep my face the same as in the original photo and make the result look natural, sharp and high quality.
Midjourney prompt
the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk, vintage 90s film photo, soft grain, slightly faded colours, nostalgic bollywood vibe, highly detailed --ar 4:5 --oref <your-photo-URL> --ow 100
Grok prompt
Edit this photo: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Vintage 90s film photo, soft grain, slightly faded colours, nostalgic Bollywood vibe. Keep my face exactly the same as in the photo. Aspect ratio 4:5, high detail.
Style 4: Royal Indian painting style mixed with realism
Royal Indian painting style mixed with realism, ornate gold details, regal mood
Nano Banana prompt
Using the uploaded photo as reference, keep every face, facial feature and identity 100% identical. Create the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Royal Indian painting style mixed with realism, ornate gold details, regal mood. Highly detailed, 4K, aspect ratio 4:5.
ChatGPT prompt
Create an image using my uploaded photo. Show the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Style: Royal Indian painting style mixed with realism, ornate gold details, regal mood. Keep the face exactly recognisable — do not change facial features, skin tone or hairline. Portrait (2:3) orientation.
Gemini prompt
Edit my photo into a new image: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Royal Indian painting style mixed with realism, ornate gold details, regal mood. Keep my face the same as in the original photo and make the result look natural, sharp and high quality.
Midjourney prompt
the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk, royal indian painting style mixed with realism, ornate gold details, regal mood, highly detailed --ar 4:5 --oref <your-photo-URL> --ow 100
Grok prompt
Edit this photo: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Royal Indian painting style mixed with realism, ornate gold details, regal mood. Keep my face exactly the same as in the photo. Aspect ratio 4:5, high detail.
Style 5: Soft dreamy editorial portrait
Soft dreamy editorial portrait, glowing bokeh, magazine-cover quality
Nano Banana prompt
Using the uploaded photo as reference, keep every face, facial feature and identity 100% identical. Create the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Soft dreamy editorial portrait, glowing bokeh, magazine-cover quality. Highly detailed, 4K, aspect ratio 4:5.
ChatGPT prompt
Create an image using my uploaded photo. Show the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Style: Soft dreamy editorial portrait, glowing bokeh, magazine-cover quality. Keep the face exactly recognisable — do not change facial features, skin tone or hairline. Portrait (2:3) orientation.
Gemini prompt
Edit my photo into a new image: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Soft dreamy editorial portrait, glowing bokeh, magazine-cover quality. Keep my face the same as in the original photo and make the result look natural, sharp and high quality.
Midjourney prompt
the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk, soft dreamy editorial portrait, glowing bokeh, magazine-cover quality, highly detailed --ar 4:5 --oref <your-photo-URL> --ow 100
Grok prompt
Edit this photo: the person in a silk saree / kurta arranging diyas on old stone temple steps by a river at dusk. Soft dreamy editorial portrait, glowing bokeh, magazine-cover quality. Keep my face exactly the same as in the photo. Aspect ratio 4:5, high detail.
How to use this prompt
🍌 Nano Banana
- Open the Gemini app or Google AI Studio and pick the 🍌 image tool (Nano Banana).
- Upload your photo (skip this for prompts that don't need one).
- Paste the prompt and generate. Ask for small tweaks in follow-up messages.
💬 ChatGPT
- Open ChatGPT and start a new chat (it uses OpenAI's GPT Image model).
- Attach your photo with the + button if the prompt needs one.
- Paste the prompt and send. If details drift, reply with a short correction.
✨ Gemini
- Open gemini.google.com or the Gemini app.
- Upload your photo if the prompt needs one.
- Paste the prompt and send, then refine with short follow-ups.
⛵ Midjourney
- Open midjourney.com (or Discord) and go to Create.
- For photo prompts, drag your photo into the prompt bar as an Omni Reference — or replace <your-photo-URL> with its link.
- Paste the prompt and generate; raise --ow (up to 1000) to keep the face or product closer.
𝕏 Grok
- Open Grok (grok.com, the app or on X) and choose Imagine / image editing.
- Upload your photo if the prompt needs one.
- Paste the prompt and generate.
Tips for better results
- Use a sharp, well-lit, front-facing photo — avoid sunglasses, heavy filters and group shots for solo prompts.
- If the face changes, reply: “Keep my face exactly the same as in the uploaded photo.”
- Change details freely — swap colours, outfits or locations to make the image yours.
- For couple or family prompts, upload one photo per person and say who is who.
- Generate 2–3 times and keep the best; small wording changes can give very different results.