Most AI video trends start from nothing. This one starts from your footage. You take a short clip of yourself standing somewhere — against a wall, by a railing, anywhere with decent light — grab its last frame, and the AI continues the scene from that exact frame: a young woman walks in from the foreground in slow motion, gathers her hair into one hand as she approaches, takes your waiting hand, twirls once under it, and ends standing face-to-face with you. It is the “Bhola Pan Saadgi” couple trend, and it looks like a film song sequence that somebody shot for you.
The whole effect is controlled by one long, strict prompt. It has to be strict: the prompt locks both faces, locks her height and posture (AI video loves shrinking the woman or leaning her onto the man’s chest — this prompt forbids both), freezes your pose until she enters, and choreographs every second of the 10-second shot. Copy it exactly as given.
Upload @image1 first, then @image2, and paste the prompt untouched. Do not shorten it — the height, posture and reaction rules are what separate this trend’s good results from the melted ones.
BHOLA PAN SAADGI COUPLE TREND — GOOGLE FLOW PROMPT (Upload: @image1 = the last frame of your clip, @image2 = the girl’s photo. 10 seconds.) Create a highly realistic, cinematic romantic video, 10 seconds long, ONE single continuous shot with no cuts, in the exact aspect ratio of @image1. === TOP PRIORITY RULES (THESE OVERRIDE EVERYTHING ELSE) === RULE 1 — FACE IDENTITY LOCK (BOTH PEOPLE, ZERO CHANGE): THE MAN — his face is a 100% exact copy of the man in @image1 in EVERY frame, first to last: same face shape, forehead, eyes, eyebrows, nose, lips, jawline, beard/moustache (or no beard), skin tone, skin texture, marks and hairline. Smiling, turning his head, holding her hand, looking at her — it is always clearly the SAME person as in @image1. Do not make him younger, slimmer, fairer, more handsome or different in any way. THE YOUNG WOMAN — her face is a 100% exact copy of the woman in @image2 in EVERY frame: same face shape, eyes, eyebrows, nose, lips, jawline, skin tone and marks. When her face appears during the twirl and at the end, she must be instantly recognisable as the woman in @image2. For both: no beautifying, no face swap, no morphing, no generic AI face, no drifting to a different face at any moment of the walk, the twirl or the face-to-face ending. RULE 2 — ADULT WOMAN + FIXED HEIGHT (STRICT): the young woman is a fully grown ADULT WOMAN, about 25 years old, with a mature adult face and normal adult body proportions, exactly as grown-up as she looks in @image2. She must NEVER look like a child, a kid or a teenager in any frame. FIXED HEIGHT: she is 5 feet 7 inches tall (170 cm); with her block heels she stands about 5 feet 9 inches. The man is about 5 feet 10 inches (178 cm), so she is only a little shorter than him. When she stands next to him, the TOP OF HER HEAD is at the level of his FOREHEAD (between his eyebrows and his hairline), her eyes just below his eyes, her shoulders only slightly lower than his. This is fixed: at NO moment is her head at his chin, neck, shoulder or chest level, and she is never taller than him. POSTURE LOCK: whenever she is next to him, she stands FULLY UPRIGHT — straight back, straight neck, knees straight. She does NOT lean toward him, does NOT bend forward, does NOT rest her head on his chest, does NOT bend her knees. He also stands straight and does not bend down to her. DEPTH LOCK: when she reaches him, her feet are on the SAME line on the floor as his feet (same distance from the camera), so their real height difference reads correctly. RULE 3 — ENVIRONMENT LOCK: the location, background, buildings, wall, railing, floor, sky, light, colours and time of day stay EXACTLY as in @image1 for all 10 seconds — nothing appears, disappears or changes. === END OF TOP PRIORITY RULES === FIRST FRAME — @image1 IS THE STARTING FRAME: the video begins exactly on @image1. Copy @image1 1:1 — same location, background, walls, railing, sky, ground, lighting, time of day, colours, camera height, camera angle, lens, framing and distance. The man in @image1 stays exactly as he is: same face, hairstyle, clothes, shoes, body, standing position and spot. Do NOT invent a new location and do NOT change the lighting or time of day — the video is a direct continuation of @image1. THE MAN (from @image1): exactly the same person (face locked as in RULE 1) — same face, age, skin tone, hairstyle, outfit and colours; do not redesign or restyle him. MAN’S POSE FREEZE — VERY IMPORTANT: from 0.0s until the young woman ENTERS the frame (about 1.0s), the man holds EXACTLY his @image1 pose like a living statue. Both hands and arms stay EXACTLY where they are in @image1 — if a hand rests on a wall, railing, pocket, waist or anything else, it STAYS resting there; do not drop, lower, lift or move his hands or arms at all. Same body posture, same lean, same feet position, same head angle, same gaze direction, same facial expression. Only tiny natural life signs are allowed: soft breathing, a natural blink, hair and clothes moving slightly in the breeze. He does NOT move or change his stance before she appears. MAN’S HERO REACTION — THE MOMENT SHE ENTERS (from about 1.2s): the instant the young woman enters the frame, he reacts IMMEDIATELY, not late. 1.2–2.0s: his eyes catch her, his face lights up with a big, charming, confident hero smile (like a Bollywood hero seeing his love), he lifts his hands off the wall / out of their @image1 position and straightens up, standing tall and proud. 2.0–3.5s: he gives a small playful head tilt and a light chin-up nod toward her, then confidently reaches ONE hand out toward her, palm up, inviting her, while she walks toward him in slow motion with the hair gesture. He stays on his spot (at most one small step forward), keeps smiling, eyes locked on her, and keeps that hand held out, waiting, until she places her hand in his at about 4.8s. His reaction must start right when she enters — he must NOT keep his hands on the wall while she walks all the way to him. THE YOUNG WOMAN (from @image2) — 100% FACE + HAIR IDENTITY LOCK: she is exactly the same person as in @image2 — same face, facial features, skin tone, same age (an adult, exactly as old as in @image2), adult body, height fixed at 5 ft 7 in as in RULE 2. HAIRSTYLE LOCK: her hair keeps the exact same length, colour, texture, parting and style as in @image2 — not longer, shorter, straighter, curlier or a different colour. FIXED WESTERN OUTFIT (IGNORE THE CLOTHES IN @image2): whatever she wears in @image2 — kurta, suit, saree, dupatta, jeans or anything else — is NOT used. She always wears this exact stylish WESTERN outfit. TOP: a fitted, solid dark chocolate-brown ribbed knit one-shoulder top (one shoulder bare, the other covered), cropped to the waist and neatly tucked in; no dupatta, no kurta, no long tunic, no scarf. TROUSERS: high-waist, wide-leg, flowy formal trousers in cream / off-white, full length down to the ankles, with a thin dark-brown leather belt at the waist; NOT jeans, NOT leggings, NOT salwar. FOOTWEAR: nude-beige block-heel sandals. ACCESSORIES: minimal — small stud earrings only. This outfit stays exactly the same in every frame: same colours, same fit, same design. HEIGHT REMINDER: fixed as in RULE 2 — she is 5 ft 7 in, the top of her head at his forehead; at the end she lifts her eyes only slightly to look into his. SPEED: the whole clip plays in smooth, gentle cinematic slow motion (about 0.6x real speed) — graceful and dreamy, while every movement stays natural and continuous, never frozen or jerky. CAMERA: completely static, locked in the exact position of @image1 — no zoom, no pan, no tilt, no push-in, no shake, no angle change; the same framing for the entire 10 seconds. ACTION TIMELINE: 0.0–1.0s — exactly @image1; the young woman is not yet visible and the man keeps his exact @image1 pose, hands and arms in place (still resting on the wall if they are there in @image1), same stance, same expression; only breathing and a blink, nothing else moves. 1.0–3.5s — SLOW-MOTION ENTRY FROM THE FOREGROUND: the young woman enters from the immediate foreground, from the bottom edge of the frame, very close to the camera, with her BACK to the camera. At first only her shoulders, hair and back fill the near foreground (slightly soft because she is so close); then she walks away from the camera, straight toward the man, coming into full sharp focus — seen from behind, her hair from @image2 and her brown one-shoulder top + cream wide-leg trousers clearly visible. 1.5–4.8s — HAIR GATHER & HOLD (THE KEY MOMENT): as she starts walking away from the camera in slow motion, she raises her RIGHT hand up behind her head, elbow pointing up and out to the side. With that hand she grabs ALL of her hair together in one bunch at the back of her head, near the nape of her neck, and pulls the whole bunch across to one side so it hangs over her right shoulder. This uncovers the back of her neck and her upper back / shoulder line, clearly visible to the camera. She KEEPS HOLDING the gathered hair in her hand with her elbow raised while she slowly walks 3–4 steps toward him — a confident, stylish, slow pose, slightly turning her head to the side. Her left arm hangs relaxed at her side. This is NOT a quick finger-comb or a light flick: the hand clearly holds the full bunch of hair to one side for about 3 seconds. At 4.8s she lets the hair go and brings her hand down softly; the hair falls naturally over her shoulder. Her hair keeps exactly the same length, colour, texture and style as in @image2 throughout. 4.8–5.5s — HAND REACH: she reaches him and places her hand in his already-outstretched hand; he holds it gently and raises it slightly above her shoulder height, still smiling like a hero. 5.5–8.0s — ONE 360° TWIRL: holding his raised hand, she performs ONE complete, slow, graceful 360-degree twirl in place, turning under their joined hands. During the turn her face becomes visible — a happy, shy smile. Her hair and outfit move naturally with the rotation. Both of her feet stay on the ground, with natural foot placement and perfect balance. 8.0–10.0s — FACE TO FACE (HOLD): after the twirl she stops facing him at about half an arm’s length — she does NOT step in close and does NOT lean in. They stand face-to-face in profile / three-quarter view to the camera, both fully upright, feet on the same floor line, holding both hands gently at waist height. The top of her head is at his forehead level; she looks just slightly up into his eyes and he looks slightly down into hers, with warm, natural smiles. Hold this romantic final pose until the end. REALISM RULES: photorealistic, like real footage filmed on a phone camera in the same light as @image1; natural human walking, natural hand contact, realistic fingers (5 per hand), realistic body rotation and foot placement; both characters keep their feet on the ground at all times — no lifting, no floating, no jumping; detailed natural skin texture, soft natural shadows matching the light in @image1, premium 4K quality; natural hair physics and realistic fabric movement; exactly two people in the whole video — the man from @image1 and the young woman from @image2. AUDIO: no dialogue, no voice, no music — only very soft natural ambient sound and footsteps (the music is added later in editing). STRICT NEGATIVE INSTRUCTIONS: do not change the location, background, lighting, time of day, camera angle, framing or aspect ratio of @image1. No camera movement, no zoom, no cuts, no transitions. Do not change the man’s face in any frame — no different face, no younger/fairer/slimmer version, no beard change — and do not change his hair or clothes. Do not move the man’s hands, arms, stance or expression BEFORE the young woman enters the frame; but once she enters, he must react immediately (smile, hands off the wall, reach out to her) — no late reaction, no staying frozen while she walks to him. He does not walk across the frame. Do not change the young woman’s face, hairstyle, hair length or hair colour from @image2. Do NOT copy her clothes from @image2 — no kurta, no suit, no salwar, no dupatta, no saree, no jeans, no ethnic wear; only the fixed western outfit (brown one-shoulder top + cream wide-leg trousers + belt + nude block heels). No outfit changes during the clip. The young woman must NEVER look like a child, kid or teenager — she is a grown adult woman, 5 ft 7 in tall. Her head is never at his chin, neck, shoulder or chest level, and never taller than him; the top of her head stays at his forehead level when she stands next to him. No leaning toward him, no bending forward, no head on his chest, no bent knees, no stepping in too close at the end. No face change for either person — no different face, no morphing, no beautification. No quick hair flick instead of the full hair-gather-and-hold gesture. Do not make either person younger — both are adults exactly as in their images. No extra people, no crowd, no extra limbs, no distorted hands or faces, no dancing, no exaggerated movements, no floating feet, no frozen frames, no text, no logos, no watermark, no artificial CGI look. FINAL RESULT: one smooth, realistic, slow-motion romantic shot that continues directly from @image1 — the young woman from @image2 (her face and hairstyle), in a stylish western outfit of a brown one-shoulder top and cream wide-leg trousers, walks in from the foreground with her back to the camera, gathers all her hair in one hand behind her head and holds it over one shoulder while walking, releases it, takes his hand, twirls once, and ends standing face-to-face with him, holding hands and smiling.
Flow is Google’s AI video creation tool. Instead of recording every scene with a camera, you give it visual references — here, a starting frame and a photo — plus a written prompt describing exactly what should happen, and it generates realistic video from them. Its strength for trends like this one is control: because the prompt can lock faces, outfits, camera position and a second-by-second timeline, the output follows a choreography instead of improvising one. With a good starting frame, a clear photo and the full prompt pasted untouched, even a first-time user can produce a clip that looks shot, not generated.
Three things feed it. The continuation idea feels like magic — viewers recognise the man’s real clip and then watch it turn into a film scene. The choreography is pure romance shorthand: the slow walk, the hair gather, the single twirl. And it is personal in a way template trends are not, because both people in the video are real people the viewer knows. Add a love song peaking at exactly the twirl, and the Reel practically shares itself.
One frame of yours, one photo of her, and the strictest prompt on this site — that is the whole production. Screenshot the last frame carefully, paste the prompt without trimming it, and the walk, the hair gather and the twirl all land where they should.
