Make your Diwali Reel look like a film scene: a row of diyas lighting up, a sparkler trail in slow motion, marigold petals falling, a dupatta flying on a palace balcony and a calm Lakshmi Puja aarti. These five cinematic AI video prompts are made for girls and women, and every scene is different.
Every example was made with AI from one picture of our fictional model Tanvi. Your own photo gives the face. Each look has two prompts: one to make a festive start photo from your selfie, and one to turn that photo into a 5-second video.
How to use this page: the examples are videos: each one plays on its own as you scroll to it (tap ▶ if it does not start). Every look has two prompts. Step 1 is a photo prompt: video apps need a photo to start from, so it turns your selfie into a festive start photo. Step 2 is the video prompt that brings that photo to life. Already have a good festive photo? Skip Step 1 and go straight to Step 2.
How it works
- Copy the photo prompt (Step 1) and use it in Gemini or ChatGPT with one clear photo of yourself.
- Check the new photo: your face, the hands and the props. Make it again if anything looks wrong; a good start photo is half of a good video.
- Open a photo-to-video tool (for example the video option in Gemini, or Kling), upload the new photo and choose 9:16 and about 5 seconds if the app asks.
- Paste the video prompt (Step 2), make two or three versions and keep the one where your face looks most natural.
- Add a festive song in Instagram and post it as a Reel.
How these video prompts are written
Each video prompt follows the same plan, so you can also write your own:
- Scene: what is already in the start photo (outfit, place, props), so the AI does not invent new things.
- Timing: what happens in seconds 0-2, 2-4 and 4-5. One small action per step keeps the face steady.
- Camera: one slow move only: a push-in, a glide, a half-circle or a backward track.
- Light and style: the mood and the cinematic film look.
- Keep and avoid: keep your face, hair and outfit the same; avoid extra people, warped hands, a flickering face and text.
1. Lighting a Row of Diyas on the Terrace
A calm opener for a Diwali Reel: one diya catches the flame and the camera slides along the row to you.
Step 1: make the start photo
✦ AI PHOTO PROMPTUpload one clear photo of yourself
Edit my uploaded photo. If I attached more than one photo of myself, use all of them together to get my face right. Do not redraw my face - keep it exactly as in the original: the same face shape and width, the same cheeks, the same jawline and chin, the same nose shape and width, the same eye shape and spacing, the same eyebrows, the same lip shape and the same natural skin tone. Do not slim my face, do not sharpen the jaw, do not enlarge the eyes, do not change my age and do not make my features more symmetrical. Change ONLY the clothing, the background and the lighting described below, and keep my head at the same angle as in the photo. Put me in an ivory and gold chikankari anarkali with a sheer gold dupatta, with gold jhumka earrings and a small red bindi. Pose: standing at a terrace parapet at night, bending slightly to light a row of clay diyas on the parapet with a long matchstick held at waist height, face turned toward the camera with a soft smile. Background: a city skyline full of warm lights and a few distant fireworks, softly out of focus. Lighting: warm diya glow on my face and outfit, deep blue night sky. Waist-up shot at eye level. Shot like a professional festive photoshoot: 85 mm portrait lens look, shallow depth of field, rich but true-to-life colours, photorealistic detail in the fabric, jewellery and skin. Vertical 9:16 format. No text, no watermark. Skin quality and face: keep the original facial identity, structure, proportions, expression and natural skin tone. Render the skin clean, flawless, healthy and naturally radiant, with an even tone across the forehead, cheeks, nose and chin. Remove AI-generated patchiness, blotches, unwanted discoloration, rough rendering, artificial noise and uneven skin texture. Make the skin high-resolution and photorealistic, with subtle natural pores, soft realistic highlights and smooth tonal transitions, refined but never plastic, waxy, blurred or over-airbrushed. The face must be sharp, clear, evenly lit and professionally photographed: no patchy complexion, muddy texture, oversharpening, grain or compression artifacts on the skin. Keep the face fully visible: no hand, arm, hair, prop or flame covers the forehead, eyes, nose, mouth or chin. Final quality: premium professional portrait photography, natural flawless skin, crisp facial detail and a realistic photographic finish.
🎨 Customize prompt size, colour, glasses
Size
Outfit colour
Keep the same
Step 2: turn it into a video
✦ AI VIDEO PROMPTUpload the start photo
Animate my photo into a 5-second cinematic vertical video (9:16). Scene: I stand at a terrace parapet at night in an ivory and gold chikankari anarkali with a sheer gold dupatta, a row of clay diyas on the parapet and a city skyline full of lights behind me. 0-2 s: I bend slightly and touch the matchstick to the next diya; its flame catches and grows. 2-4 s: I straighten up slowly, my dupatta moving in a light breeze, while small golden fireworks bloom far away in the sky. 4-5 s: I look at the camera with a soft smile and hold still. Camera: a slow, smooth glide along the row of diyas toward me, ending waist-up. Light: warm diya glow on my face and outfit, deep blue night sky, tiny fireworks in the distance. Style: cinematic film look, smooth realistic motion, shallow depth of field, 24 fps. Keep my face, hairstyle, outfit and jewellery exactly as in the photo, with natural blinking and no change to my features. Avoid: extra people, fast or jerky movement, warped hands or fingers, flickering face, text, logos and watermarks.
2. Sparkler Twirl in Slow Motion
Slow motion makes the sparkler trail look magical. Keep the sparkler at arm’s length in your photo so the AI keeps it away from your face.
Step 1: make the start photo
✦ AI PHOTO PROMPTUpload one clear photo of yourself
Edit my uploaded photo. If I attached more than one photo of myself, use all of them together to get my face right. Do not redraw my face - keep it exactly as in the original: the same face shape and width, the same cheeks, the same jawline and chin, the same nose shape and width, the same eye shape and spacing, the same eyebrows, the same lip shape and the same natural skin tone. Do not slim my face, do not sharpen the jaw, do not enlarge the eyes, do not change my age and do not make my features more symmetrical. Change ONLY the clothing, the background and the lighting described below, and keep my head at the same angle as in the photo. Put me in a deep emerald green lehenga with gold embroidery and a matching dupatta, with gold chandbali earrings and a small bindi. Pose: standing in a courtyard at night, holding a lit sparkler out to the side at arm’s length, laughing toward the camera. Background: strings of warm fairy lights and marigold garlands, softly out of focus, a few fireworks in the sky. Lighting: bright golden sparkler light on one side of my face, warm fairy-light glow. Waist-up shot at eye level. Shot like a professional festive photoshoot: 85 mm portrait lens look, shallow depth of field, rich but true-to-life colours, photorealistic detail in the fabric, jewellery and skin. Vertical 9:16 format. No text, no watermark. Skin quality and face: keep the original facial identity, structure, proportions, expression and natural skin tone. Render the skin clean, flawless, healthy and naturally radiant, with an even tone across the forehead, cheeks, nose and chin. Remove AI-generated patchiness, blotches, unwanted discoloration, rough rendering, artificial noise and uneven skin texture. Make the skin high-resolution and photorealistic, with subtle natural pores, soft realistic highlights and smooth tonal transitions, refined but never plastic, waxy, blurred or over-airbrushed. The face must be sharp, clear, evenly lit and professionally photographed: no patchy complexion, muddy texture, oversharpening, grain or compression artifacts on the skin. Keep the face fully visible: no hand, arm, hair, prop or flame covers the forehead, eyes, nose, mouth or chin. Final quality: premium professional portrait photography, natural flawless skin, crisp facial detail and a realistic photographic finish.
🎨 Customize prompt size, colour, glasses
Size
Outfit colour
Keep the same
Step 2: turn it into a video
✦ AI VIDEO PROMPTUpload the start photo
Animate my photo into a 5-second cinematic vertical video (9:16). Scene: I stand in a courtyard at night in a deep emerald green lehenga with gold embroidery, strings of fairy lights and marigold garlands behind me. 0-2 s: in slow motion I draw a slow circle in the air with the lit sparkler held out at arm’s length, leaving a glowing trail of light. 2-4 s: sparks fall softly to the ground, my chandbali earrings and dupatta sway, and fireworks burst far behind me. 4-5 s: I turn to the camera and laugh, the sparkler still held out to the side, away from my face. Camera: a slow push-in from waist-up to chest-up. Light: bright golden sparkler light on one side of my face, warm fairy-light bokeh behind. Style: cinematic film look, smooth realistic motion, shallow depth of field, 24 fps. Keep my face, hairstyle, outfit and jewellery exactly as in the photo, with natural blinking and no change to my features. Avoid: extra people, fast or jerky movement, warped hands or fingers, flickering face, text, logos and watermarks.
3. Marigold Petal Rain by the Rangoli
Falling petals add movement without changing your face much, so this one keeps the face steadier than most.
Step 1: make the start photo
✦ AI PHOTO PROMPTUpload one clear photo of yourself
Edit my uploaded photo. If I attached more than one photo of myself, use all of them together to get my face right. Do not redraw my face - keep it exactly as in the original: the same face shape and width, the same cheeks, the same jawline and chin, the same nose shape and width, the same eye shape and spacing, the same eyebrows, the same lip shape and the same natural skin tone. Do not slim my face, do not sharpen the jaw, do not enlarge the eyes, do not change my age and do not make my features more symmetrical. Change ONLY the clothing, the background and the lighting described below, and keep my head at the same angle as in the photo. Put me in a bright orange silk saree with a magenta border, with gold jhumka earrings, a gold necklace and a small red bindi. Pose: sitting on the floor beside a big colourful rangoli with lit diyas around it, hands resting in my lap, looking toward the camera with a happy smile. Background: a home courtyard with marigold garlands and warm fairy lights, softly out of focus. Lighting: warm evening light with diya glow. Seated shot from the lap up, camera slightly above. Shot like a professional festive photoshoot: 85 mm portrait lens look, shallow depth of field, rich but true-to-life colours, photorealistic detail in the fabric, jewellery and skin. Vertical 9:16 format. No text, no watermark. Skin quality and face: keep the original facial identity, structure, proportions, expression and natural skin tone. Render the skin clean, flawless, healthy and naturally radiant, with an even tone across the forehead, cheeks, nose and chin. Remove AI-generated patchiness, blotches, unwanted discoloration, rough rendering, artificial noise and uneven skin texture. Make the skin high-resolution and photorealistic, with subtle natural pores, soft realistic highlights and smooth tonal transitions, refined but never plastic, waxy, blurred or over-airbrushed. The face must be sharp, clear, evenly lit and professionally photographed: no patchy complexion, muddy texture, oversharpening, grain or compression artifacts on the skin. Keep the face fully visible: no hand, arm, hair, prop or flame covers the forehead, eyes, nose, mouth or chin. Final quality: premium professional portrait photography, natural flawless skin, crisp facial detail and a realistic photographic finish.
🎨 Customize prompt size, colour, glasses
Size
Outfit colour
Keep the same
Step 2: turn it into a video
✦ AI VIDEO PROMPTUpload the start photo
Animate my photo into a 5-second cinematic vertical video (9:16). Scene: I sit on the floor beside a big colourful rangoli in a bright orange silk saree with a magenta border, lit diyas around the rangoli and marigold garlands behind me. 0-2 s: orange and yellow marigold petals start falling slowly from above like gentle rain; I look up at them. 2-4 s: I open my palms in my lap and catch a few petals, smiling as they land. 4-5 s: I look at the camera with a happy smile while petals keep drifting down and the diya flames flicker. Camera: a slow rise from the rangoli up to my face, ending chest-up. Light: warm evening light with soft diya glow, petals catching the light as they fall. Style: cinematic film look, smooth realistic motion, shallow depth of field, 24 fps. Keep my face, hairstyle, outfit and jewellery exactly as in the photo, with natural blinking and no change to my features. Avoid: extra people, fast or jerky movement, warped hands or fingers, flickering face, text, logos and watermarks.
4. Palace Balcony With Flowing Dupatta
The wind and the half-circle camera move make this the most cinematic look in the set. Use a photo with a long dupatta.
Step 1: make the start photo
✦ AI PHOTO PROMPTUpload one clear photo of yourself
Edit my uploaded photo. If I attached more than one photo of myself, use all of them together to get my face right. Do not redraw my face - keep it exactly as in the original: the same face shape and width, the same cheeks, the same jawline and chin, the same nose shape and width, the same eye shape and spacing, the same eyebrows, the same lip shape and the same natural skin tone. Do not slim my face, do not sharpen the jaw, do not enlarge the eyes, do not change my age and do not make my features more symmetrical. Change ONLY the clothing, the background and the lighting described below, and keep my head at the same angle as in the photo. Put me in a royal blue velvet lehenga with silver embroidery and a long sheer silver dupatta, with a silver choker, silver jhumka earrings and a small bindi. Pose: standing at a carved stone palace balcony lined with lit diyas, one hand resting on the railing, the dupatta trailing behind me, looking at the camera with a calm smile. Background: an old palace courtyard lit with hundreds of diyas and lanterns at dusk, softly out of focus. Lighting: warm diya glow with a cool blue dusk sky. Three-quarter-length shot at eye level. Shot like a professional festive photoshoot: 50 mm lens look, the whole outfit in sharp focus, background softly blurred, rich but true-to-life colours, photorealistic detail. Vertical 9:16 format. No text, no watermark. Skin quality and face: keep the original facial identity, structure, proportions, expression and natural skin tone. Render the skin clean, flawless, healthy and naturally radiant, with an even tone across the forehead, cheeks, nose and chin. Remove AI-generated patchiness, blotches, unwanted discoloration, rough rendering, artificial noise and uneven skin texture. Make the skin high-resolution and photorealistic, with subtle natural pores, soft realistic highlights and smooth tonal transitions, refined but never plastic, waxy, blurred or over-airbrushed. The face must be sharp, clear, evenly lit and professionally photographed: no patchy complexion, muddy texture, oversharpening, grain or compression artifacts on the skin. Keep the face fully visible: no hand, arm, hair, prop or flame covers the forehead, eyes, nose, mouth or chin. Final quality: premium professional portrait photography, natural flawless skin, crisp facial detail and a realistic photographic finish.
🎨 Customize prompt size, colour, glasses
Size
Outfit colour
Keep the same
Step 2: turn it into a video
✦ AI VIDEO PROMPTUpload the start photo
Animate my photo into a 5-second cinematic vertical video (9:16). Scene: I stand at a carved stone palace balcony lined with lit diyas, in a royal blue velvet lehenga with silver embroidery and a long sheer silver dupatta, a palace courtyard full of diyas and lanterns behind me at dusk. 0-2 s: a strong soft wind lifts my long dupatta and my hair behind me in slow motion. 2-4 s: I turn my face from the courtyard toward the camera as the dupatta keeps flowing. 4-5 s: I smile at the camera, the diya flames on the balcony flickering and the lanterns below glowing brighter. Camera: a slow half-circle move around me, from the side to the front, three-quarter length. Light: warm diya glow against a cool blue dusk sky. Style: cinematic film look, smooth realistic motion, shallow depth of field, 24 fps. Keep my face, hairstyle, outfit and jewellery exactly as in the photo, with natural blinking and no change to my features. Avoid: extra people, fast or jerky movement, warped hands or fingers, flickering face, text, logos and watermarks.
5. Lakshmi Puja Aarti
A gentle, devotional clip for Lakshmi Puja night. Hold the thali at chest height in your photo.
Step 1: make the start photo
✦ AI PHOTO PROMPTUpload one clear photo of yourself
Edit my uploaded photo. If I attached more than one photo of myself, use all of them together to get my face right. Do not redraw my face - keep it exactly as in the original: the same face shape and width, the same cheeks, the same jawline and chin, the same nose shape and width, the same eye shape and spacing, the same eyebrows, the same lip shape and the same natural skin tone. Do not slim my face, do not sharpen the jaw, do not enlarge the eyes, do not change my age and do not make my features more symmetrical. Change ONLY the clothing, the background and the lighting described below, and keep my head at the same angle as in the photo. Put me in a red silk saree with a gold border, with a gold necklace, gold jhumka earrings, a red bindi and red and gold bangles. Pose: standing in front of a decorated home puja corner, holding a brass aarti thali with a lit diya and flowers at chest height, looking at the camera with a peaceful smile. Background: a puja mandir with marigold garlands, small brass lamps and a soft glow, softly out of focus, no idols in sharp focus. Lighting: warm golden lamp light. Waist-up shot at eye level. Shot like a professional festive photoshoot: 85 mm portrait lens look, shallow depth of field, rich but true-to-life colours, photorealistic detail in the fabric, jewellery and skin. Vertical 9:16 format. No text, no watermark. Skin quality and face: keep the original facial identity, structure, proportions, expression and natural skin tone. Render the skin clean, flawless, healthy and naturally radiant, with an even tone across the forehead, cheeks, nose and chin. Remove AI-generated patchiness, blotches, unwanted discoloration, rough rendering, artificial noise and uneven skin texture. Make the skin high-resolution and photorealistic, with subtle natural pores, soft realistic highlights and smooth tonal transitions, refined but never plastic, waxy, blurred or over-airbrushed. The face must be sharp, clear, evenly lit and professionally photographed: no patchy complexion, muddy texture, oversharpening, grain or compression artifacts on the skin. Keep the face fully visible: no hand, arm, hair, prop or flame covers the forehead, eyes, nose, mouth or chin. Final quality: premium professional portrait photography, natural flawless skin, crisp facial detail and a realistic photographic finish.
🎨 Customize prompt size, colour, glasses
Size
Outfit colour
Keep the same
Step 2: turn it into a video
✦ AI VIDEO PROMPTUpload the start photo
Animate my photo into a 5-second cinematic vertical video (9:16). Scene: I stand in front of a decorated home puja corner in a red silk saree with a gold border, holding a brass aarti thali with a lit diya and flowers at chest height, marigold garlands and small brass lamps behind me. 0-2 s: I slowly move the aarti thali in a gentle circle at chest height; the diya flame glows and casts soft light on my face. 2-4 s: small brass bells in the background sway gently; my eyes close for a moment in calm devotion. 4-5 s: I hold the thali still below my face, open my eyes and smile at the camera. Camera: a slow, steady push-in, ending chest-up. Light: warm golden lamp light, calm and devotional mood. Style: cinematic film look, smooth realistic motion, shallow depth of field, 24 fps. Keep my face, hairstyle, outfit and jewellery exactly as in the photo, with natural blinking and no change to my features. Avoid: extra people, fast or jerky movement, warped hands or fingers, flickering face, text, logos and watermarks.
Outfit ideas that look good on video
Rich colours read best on video under warm diya light: emerald, royal blue, red, orange and ivory with gold all work. Long dupattas and lehenga flares move beautifully in slow motion, so choose a flowing outfit for the balcony and sparkler looks. Heavy jhumkas and chandbalis catch the light when you turn your head.
Make it a full Reel
- Make two or three of these clips in the same outfit and join them in VN or CapCut. Three 5-second clips make a 15-second Reel.
- Cut on the beat of your song: one clip per beat drop looks planned.
- Put a slow-motion clip at the start; it makes people stop scrolling.
- Keep the original colours. A heavy filter on top of AI video often makes faces look waxy.
Before you post the video
Watch the clip twice at full size. Check the face in the last second, when the head turns to the camera, and check the hands holding the diya, sparkler or lantern. Flames and sparks should stay away from the face. If the face drifts, make the video again from the same photo; small motion keeps the face steadier.
Mistakes to avoid
- Asking for big moves like dancing, spinning or a full head turn. The face changes most during big moves.
- Changing the Scene line so it no longer matches your photo. The AI then tries to redraw the outfit or the place.
- Using a blurry or dark start photo. The video can only be as sharp as the photo.
- Holding a sparkler, diya or cracker close to your face in the start photo. Keep fire at arm’s length, in AI and in real life.
Caption ideas for your Diwali Reel
- Diyas, dupatta and a little Diwali magic ✨
- Lighting up this Diwali, one diya at a time 🪔
- Slow motion, fast heartbeat. Happy Diwali!
- Petals, rangoli and pure festive vibes.
- Shubh Deepavali from my home to yours.
Common questions
Can I skip Step 1 and use my own Diwali photo? Yes. Use the video prompt with any festive photo, and change the Scene line to describe your photo, for example “candle” instead of “diya”.
Why is my video only 5 seconds? Most photo-to-video tools make 5 to 10 second clips. Make two clips and join them in VN or CapCut for a longer Reel.
Why does my face change in the video? Big movements make the AI redraw the face. Use a sharp photo where your face is clear, and keep the motion words gentle.
Gemini or Kling? Both accept a start photo and a prompt. Free limits and quality change often, so try the same prompt in both and keep the better clip.
More ideas: the Diwali cinematic video prompts for men, the Dhanteras video prompts and the guide to keeping your face the same.
More trends
View all →
AI Video EditingDhanteras AI Video Prompts: Doorstep Diya and Gold Necklace Videos From One Photo
AI Video EditingDiwali Cinematic AI Video Prompts for Men: 5 Fireworks, Anar and Haveli Videos
AI Video EditingRoyal Diwali Cinematic AI Video Prompts: Ganga Ghat, Palace Princess and Fireworks (8-Second Videos)
AI Video EditingKarwa Chauth Moon Chalni Reel Edit Guide 2026
AI Video EditingPetrol Pump AI Dance: Gemini Guide with 3 Photos and a Reference Video
AI Video Editing