One idea → a short film
Make a movie

Prompt library · Seedance 2.5

AI video prompts for Seedance 2.5, with Sora prompts that still work

A good AI video prompt is short and physical: who moves, how, what the camera does and what you hear. Seedance 2.5 follows that order closely. These prompts are written for image-to-video, so each one starts from a still and only describes what happens next, which is the most reliable way to keep a face stable across a clip.

  • 10 tested prompts
  • Made with Seedance 2.5 + GPT Image 2.5
  • Real outputs, unedited

Clips marked "Seedance 2.5 Turbo" are real 5-second 720p renders made on MakeInfluencer.ai. Every start frame is a GPT Image 2.5 still of our own AI character from our other prompt pages. MakeInfluencer.ai is not affiliated with ByteDance or OpenAI. OpenAI retired the Sora API on 24 September 2026; the Sora-style prompts below are rewritten for Seedance 2.5.

Woman at a café window table lifting a coffee cup and laughing
Woman by a rainy window turning from the rain toward the camera with a small smile
Woman in a white linen dress walking through tall grass at sunset

Seedance 2.5 image-to-video prompts

Calm, identity-safe motion for social clips. Start from a sharp still at the aspect ratio you want to post.

No. 01Seedance 2.5 Turbo · 720p clip

Café coffee sip

She lifts the cup, takes a small sip, lowers it and laughs softly at something off camera to the left. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Camera nearly static, a very slow push-in. Natural, believable motion: steam from the cup, people passing outside the window out of focus. Sound: quiet café ambience, cups clinking, no music. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic candid photograph of the woman sitting at a small marble café table by the window in late morning. She wears a cream cable-knit sweater with pushed-up sleeves and small gold hoop earrings, holds a cup of coffee in both hands and looks off to the left with a relaxed half-smile, as if a friend just said something. A croissant on a small plate in front of her; the street outside softly out of focus. Framed from mid-chest up, slightly off-centre, eye level, as if taken by a friend on a phone. Soft daylight from the window, gentle shadow on the far side of her face. Natural phone-photo colour. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Café window candid

Upload one clear photo of yourself, then paste the prompt.

No. 02Seedance 2.5 Turbo · 720p clip

Rainy window glance

Rain keeps running down the glass. She watches it for a moment, then slowly turns her head toward the camera, no further than three-quarter view, and a small smile reaches the corners of her mouth. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Static camera. Subtle motion only: breathing, blinking, steam from the mug. Sound: steady rain on the window, a distant car, no music. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic moody portrait of the woman sitting by a large window on a rainy afternoon, wrapped in an oversized cream knit sweater, holding a ceramic mug in both hands. She looks out at the rain with a thoughtful expression, face in three-quarter view. Raindrops and streaks on the glass in the foreground, slightly out of focus, blurred city lights beyond. Framed from the chest up, shot through the glass at eye level. Soft grey daylight from the window lighting her face, the room behind her falling into shadow. Muted cool colour grade, gentle film grain. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Rainy window portrait

Upload one clear photo of yourself, then paste the prompt.

No. 03Seedance 2.5 Turbo · 720p clip

Golden hour walk

She keeps walking slowly away through the tall grass, her hand brushing the grass tops, then glances back over her shoulder with a soft smile. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Camera follows gently at walking pace. The linen dress and her hair move in the breeze; grass sways; low sun flares softly. Sound: wind in the grass, crickets, footsteps, no music. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic golden-hour photograph of the woman walking through a field of tall dry grass, wearing a loose white linen dress with thin straps. She glances back over her shoulder with a soft smile, one hand trailing through the grass tops. The sun is low behind her, creating a warm rim light around her hair and shoulders and a gentle lens flare. Framed from the knees up, slightly below eye level, 85mm look with the background melting into warm bokeh. The linen shows natural wrinkles and moves in the breeze. Warm, honest colour, not oversaturated. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Golden hour field

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman holding a dropper bottle in a bathroom selfieNo. 04Start frame · GPT Image 2.5

UGC talking selfie

Dialogue language: American English. She looks into the lens and says, relaxed and a little amused: {Okay, two weeks in and my skin actually feels calmer.} She tilts the bottle toward the camera, then back. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Handheld selfie framing with slight arm movement, natural phone-camera look. Sound: her voice, small bathroom room tone, no music, no other spoken words. Natural blinking, no exaggerated expressions. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create an unedited front-camera phone selfie of the woman in her bathroom in the morning, filming a quick skincare recommendation. Nearly bare face, hair pulled back with a claw clip and a few loose strands. She holds a small frosted-glass dropper bottle with a white cap and no label up beside her cheek, fingers wrapped naturally around it, and smiles mid-sentence looking into the lens. Shoulders up at arm's length, slight wide-angle distortion. Daylight from a frosted window on her left, white tiles and a towel hook behind her. Phone camera look: slight noise, natural colour. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: UGC product selfie

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman holding mulled wine at a Christmas marketNo. 05Start frame · GPT Image 2.5

Christmas market snowfall

Snow keeps falling. She raises the mug, blows gently across the top so the steam drifts, then smiles at the camera. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Very slow push-in. String lights twinkle softly behind her, a few blurred shoppers pass. Sound: market chatter, a distant brass band, cups clinking. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic photograph of the woman at a European-style Christmas market at dusk, holding a steaming ceramic mug of mulled wine in gloved hands and smiling at the camera. She wears a camel wool coat, a cream knit scarf and a red beanie. Wooden stalls with warm string lights and pine garlands behind her, softly out of focus, no readable signs. Light snow falling. Framed from the waist up, eye level, 50mm look. Warm stall lights on her face against the deep blue evening sky. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Christmas market at dusk

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman dressed as a witch holding a lantern in a foggy forestNo. 06Start frame · GPT Image 2.5

Witch with a lantern

The candle in the lantern flickers. She slowly raises the lantern to face height, the warm light moving across her face, and her half-smile widens slightly. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Static camera. Fog drifts between the trees, a few leaves fall. Sound: wind through bare branches, a distant owl, leaves rustling, no music. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic Halloween portrait of the woman as a witch in a misty autumn forest at dusk. She wears a wide-brimmed black felt witch hat, a long black wool cloak over a dark green dress and holds an old brass lantern with a candle inside. Fallen orange leaves on the ground, bare trees fading into fog behind her. She looks at the camera with a calm, knowing half-smile. Framed from the waist up, eye level, 50mm look. Warm candlelight from the lantern on her face, cool blue fog light behind. Fine film grain. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Witch in a misty forest

Upload one clear photo of yourself, then paste the prompt.

Sora prompts that work in Seedance 2.5

Sora-era prompts were long cinematic paragraphs. Seedance 2.5 wants the same ideas in its order: subject and action, scene, camera, sound. These are rewritten that way.

No. 07Seedance 2.5 Turbo · 720p clip

Neon street, cinematic

She stands on the wet street as light rain begins, looks up at the sky for a beat, then back toward the camera with a slight smile. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Slow push-in from medium to medium close-up. Neon reflections ripple in the puddles, rain streaks catch the pink and teal light. Style: anamorphic film look, deep blacks, fine grain. Sound: rain on pavement, distant traffic, a faint synth hum. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic cinematic night photograph of the woman standing on a narrow wet city street, wearing a black leather jacket over a white T-shirt. She looks just past the camera, hands in her jacket pockets, a few strands of hair across her face in the breeze. Blurred shop signs glow pink and teal behind her with no readable lettering, reflections on the wet asphalt. Framed from the waist up, eye level, 35mm look, shallow depth of field. Neon light from both sides: magenta on one cheek, cyan on the other. Slight haze, fine grain, deep blacks. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Neon night street

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman in a camel coat crossing a city streetNo. 08Start frame · GPT Image 2.5

Street-style tracking shot

She continues across the crosswalk at an easy pace, coat moving with each step, and glances at the camera as she reaches the kerb. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Tracking shot: the camera moves sideways at her walking speed, she stays sharp while the street slides behind her. Walking carries real weight. Sound: city traffic, footsteps, a bicycle bell. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic editorial street-style photograph of the woman walking across a city crosswalk in late afternoon, mid-stride, glancing just past the camera. She wears an oversized camel wool coat over a white ribbed knit top, straight-leg dark jeans and white leather trainers, and carries a plain black leather shoulder bag with no visible logo. Full body visible, feet included, shot from slightly below eye level, 35mm look, pedestrians and buildings softly out of focus. Low warm sun from the side, long shadows on the asphalt. 35mm film look, subtle grain. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Street-style crosswalk

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman with a clear umbrella in a Kyoto lane at dawnNo. 09Start frame · GPT Image 2.5

Kyoto lane at dawn

She turns forward and walks slowly up the lane, twirling the clear umbrella once. Lantern light flickers on the wet stones. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Camera holds, then slowly cranes up to reveal the pagoda roof in the haze. Style: 35mm film look, cool dawn blue with warm lantern glow. Sound: dripping water, soft footsteps, distant temple bell. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic travel photograph of the woman in a narrow Kyoto lane with wooden machiya houses at dawn after rain. She wears a beige trench coat, holds a clear umbrella and looks back over her shoulder at the camera, medium distance. Wet stone paving in the foreground reflecting paper lanterns, houses in the midground, a pagoda roofline in soft haze behind. 35mm look, eye level. Cool blue dawn light with warm lantern glow. Film look with fine grain. No other people. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Travel photo

Upload one clear photo of yourself, then paste the prompt.

Start frame of a woman in a red saree on a rooftop at golden hourNo. 10Start frame · GPT Image 2.5

Saree in the breeze

A warm breeze lifts the end of the red saree; she lets it trail from her hand, closes her eyes for a moment and smiles as her hair moves. Keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Slow orbit a few degrees to the right, no more. Silk ripples and catches the low sun. Sound: wind, distant city, birds, no music. No on-screen text, no subtitles, no logos, no morphing or face warping.

Start frame prompt

Use the uploaded photo only for the woman's identity. Keep her face exactly as in the photo: face shape, eyes, eyelids, eyebrows, nose, lips, jawline, hairline, hair colour and skin tone. Do not beautify, slim, smooth or retouch her face; keep its natural asymmetry and texture. Create a photorealistic candid photograph of the woman on a flat city rooftop at golden hour, wearing a deep red silk saree with a narrow gold zari border and small gold jhumka earrings. A string of white jasmine is pinned into a low bun. She holds the end of the pallu in one hand as the breeze lifts it, looking off to the side with a relaxed smile. Water tanks and a hazy skyline soft in the background. Framed from mid-thigh up, eye level, 35mm look. Low warm sun from behind her right shoulder creating a rim light on her hair and the silk. The silk shows natural creases and sheen. Real skin texture: visible pores, fine vellus hair, slightly uneven natural tone. Unretouched, no skin smoothing, no beauty filter, no glamorization. No text, no logos, no watermark.

See the still: Red saree at golden hour

Upload one clear photo of yourself, then paste the prompt.

Why these prompts work

Three rules behind every prompt on this page

  1. 01

    Describe what happens next

    In image-to-video, the frame already shows the person and the place. Re-describing them wastes words; describe the motion that starts from the frame.

  2. 02

    One camera move per shot

    Slow push-in, or a tracking shot, or a static camera. "Push in while orbiting and tilting" is where clips fall apart.

  3. 03

    Write the sound and the exclusions

    Seedance 2.5 generates audio in the same pass. Name the ambience, then rule out music, subtitles and text. There is no negative prompt field.

How to write your own

How to write your own AI video prompt

Seedance 2.5's own manual gives one formula: subject and action first, then scene, style, camera and sound. For image-to-video, trim it further.

  1. 1

    Make a strong start frame

    Generate a sharp still at the aspect ratio you want to post, with the face large enough to read. The clip inherits its identity and framing.

  2. 2

    Describe one action

    What happens next, in one or two sentences. One main event per shot; two events compete for the motion budget.

  3. 3

    Lock identity

    Add: keep her face, hair, wardrobe, set and lighting exactly as in the start frame. Keep head turns to three-quarter view.

  4. 4

    Pick one camera move

    Static, slow push-in, tracking or a small orbit. Say what it does to the frame.

  5. 5

    Write sound and exclusions

    Name the ambience and any dialogue in braces, then add no music, no subtitles, no on-screen text, no morphing.

FAQ

Frequently Asked Questions

Run these prompts on your own photos

Sign up to use GPT Image 2.5, Nano Banana Pro and Seedance 2.5 in one studio, with your AI character saved for every shot.

Start creating in minutes · Pay as you go