24 Vidu Prompts for Reference-to-Video (2026)

Reference-to-video is Vidu's signature mode: upload up to seven reference images — characters, props, scenes, even unrelated ones — and Vidu fuses them into one cohesive clip while keeping each subject looking like its reference. It runs on the Vidu Q2 line from Shengshu Technology, at 1080p and 24fps with synced 48kHz audio. The whole skill is in the prompt: name which image each element comes from, describe one clear interaction, then end with an identity-lock line like "keep each character consistent with its reference image."

Every prompt below follows that shape and carries a Settings note — reference count, aspect ratio, length and movement amplitude are app controls in Vidu, not inline parameters, so set them there and swap the bracketed parts. These 24 are all new scenarios; for the wider set across every mode see the full Vidu prompts roundup, and keep the Vidu prompt cheat sheet open for the camera-move and amplitude tables.

Consistent characters (1 reference)

Start with the simplest reliable setup: one locked character, a simple background, and one small action. These keep a face true from clip to clip so you can build a series. Hold amplitude at small or medium so the likeness never drifts.

1. Character turns to camera

Using the reference image, the character stands in a plain softly-lit studio and slowly turns their head toward the camera with a calm, natural expression, a slight smile forming, slow push-in on the face, soft even portrait light, clean neutral background, keep the character's face, hair and outfit consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 9:16 · 1080p · 5s · amplitude: small. Why it works: a single small head turn on a plain set gives Vidu almost nothing to distort, so the reference face stays locked.

2. Character laughs and waves

Using the reference image, the character breaks into a warm genuine laugh and raises one hand to wave at the camera, shoulders relaxing, gentle handheld feel with a slow ease-in, bright friendly daylight, simple blurred background, keep the character's face, hair and outfit exactly consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 9:16 · 1080p · 5s · amplitude: medium. Best for: creator intros and channel bumpers that reuse the same host every time.

3. Character sips coffee at a desk

Using the reference image, the character sits at a tidy wooden desk, lifts a mug of coffee, takes a slow sip and sets it back down while glancing at the camera, soft morning window light, cozy home-office background slightly blurred, slow subtle push-in, keep the character's face, hair and outfit consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: small. Why it works: one contained hand-to-mug action keeps the body still, so Vidu holds the face steady through the motion.

4. Character reacts with surprise

Using the reference image, the character's eyes widen and eyebrows lift into a delighted surprised expression, one hand rising toward their mouth, a quick natural head recoil, punchy reaction-shot framing, bright clean light, plain background, slow push-in, keep the character's face, hair and outfit exactly consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 9:16 · 1080p · 4s · amplitude: medium. Best for: reaction cutaways and thumbnails that need the same recognizable character.

Character in a new scene

Keep the same one-character reference, then describe a fully new world around it. Vidu drops the locked subject into the setting you write while holding the face true. Match a camera move to the mood and keep the environment described simply.

5. Character in a snowy forest

Using the reference image, the character walks slowly through a quiet snow-covered pine forest, breath misting in the cold air, snowflakes drifting down and catching on their coat, slow tracking shot moving beside them, soft blue-white winter light, calm and cinematic, keep the character's face, hair and outfit consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: medium. Why it works: the drifting snow supplies secondary motion while the character's slow walk keeps the locked face readable.

6. Character on a rainy neon street

Using the reference image, the character stands under a glowing storefront on a rainy city street at night, neon signs reflected in the wet pavement, they look up as rain falls, slow push-in with shallow depth of field, moody magenta-and-teal light, keep the character's face, hair and outfit exactly consistent with the reference image, soft rain sound.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: medium. Best for: cinematic character posters and moody vertical cutdowns of a recurring hero.

7. Character on a sunlit beach

Using the reference image, the character strolls along the edge of a bright sunlit beach, small waves rolling in beside them and wind moving their hair, they smile out toward the sea, slow tracking shot easing alongside, warm golden afternoon light with a gentle lens flare, keep the character's face, hair and outfit consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: medium. Why it works: a single alongside track plus rolling waves gives the scene life without asking the character to do anything that risks the face.

8. Character in a cozy library

Using the reference image, the character browses a tall shelf in a warm old library, pulls out a book and opens it thoughtfully, dust motes floating in shafts of amber lamplight, slow orbit from the side to the front, quiet studious mood, rich wood tones, keep the character's face, hair and outfit consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: small. Best for: narrative and book-related content that reuses one protagonist across settings.

Two- and three-character interactions

Give each character its own reference image, name them explicitly so they do not merge, and write one shared action between them. A single interaction — a toast, a handoff — is what Vidu solves cleanly. Medium amplitude suits full-body movement here.

9. Two friends toast at a bar

Using the reference images, the friend from image 1 and the friend from image 2 sit at a warm dimly-lit bar, raise their glasses and clink them together in a happy toast, laughing, slow push-in between the two of them, cozy amber light and soft bokeh behind, keep both characters consistent with their reference images, warm ambient chatter.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 16:9 · 1080p · 6s · amplitude: medium. Why it works: one shared action — the clink — gives Vidu a single interaction to solve between two named, locked faces.

10. Mentor hands a scroll to a student

Using the reference images, the elder mentor from image 1 hands a rolled scroll to the young student from image 2 in the temple courtyard from image 3, the student receiving it with both hands and a respectful nod, slow orbit around the pair, soft morning light, calm reverent mood, keep both characters consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 3 reference images · 16:9 · 1080p · 6s · amplitude: medium. Best for: story beats that fuse two characters and a scene reference into one shot.

11. Three bandmates on a rooftop

Using the reference images, the three bandmates from images 1, 2 and 3 stand together on a city rooftop at golden hour, turning toward the camera and grinning as the wind moves their hair and jackets, slow arcing tracking shot across the group, warm backlight and hazy skyline, keep each character consistent with its reference image.

Settings: Vidu Q2 Reference-to-Video · 3 reference images · 16:9 · 1080p · 6s · amplitude: medium. Why it works: naming images 1, 2 and 3 keeps three faces separated, and one shared turn gives the group a single unified action.

12. Detective and witness in a diner

Using the reference images, the detective from image 1 leans across a diner booth table listening intently as the nervous witness from image 2 speaks and gestures, coffee cups between them, slow push-in over the detective's shoulder, moody late-night light through the window blinds, tense quiet mood, keep both characters consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 16:9 · 1080p · 6s · amplitude: medium. Best for: dialogue-driven scenes where two consistent characters must share one framed interaction.

Product & brand references

Upload the product as one reference and, if you have it, the scene as another. Describe how the product sits or is handled, keep the background clean, and lock the packaging. Hold amplitude at small so the label never warps.

13. Hand places the bottle on a shelf

Using the reference image, a hand gently places the product bottle from the reference onto a clean minimalist shelf and withdraws, the bottle settling into a soft pool of studio light, slow push-in on the label, bright premium commercial lighting, neutral background, keep the product packaging and label exactly consistent with its reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 1:1 · 1080p · 5s · amplitude: small. Why it works: a slow push-in on a locked label with almost no camera motion keeps the real packaging crisp and readable.

14. Headphones on a studio desk

Using the reference images, the headphones from image 1 rest on the wooden studio desk from image 2 beside a laptop and a warm lamp, a soft light sweep passing across the ear cups, slow dolly across the desk, cozy modern tech-lifestyle mood, shallow focus, keep the headphones exactly consistent with their reference image.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 16:9 · 1080p · 5s · amplitude: small. Best for: gear and gadget ads that drop a real product into a styled desk scene.

15. Sneaker on a gym floor

Using the reference image, the sneaker from the reference sits on a polished gym floor under a shaft of hard directional light, a slow low orbit circling the shoe to show every angle, faint drifting dust in the beam, high-contrast athletic commercial grade, dark background, keep the sneaker's shape, colorway and logo exactly consistent with its reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 1:1 · 1080p · 5s · amplitude: small. Why it works: a slow low orbit reveals the product while low amplitude stops the silhouette and logo from morphing.

16. Skincare jar on a bathroom counter

Using the reference images, the skincare jar from image 1 stands on the clean marble bathroom counter from image 2 in soft morning light, a folded towel and a sprig of eucalyptus beside it, slow gentle push-in on the jar, calm spa-like mood, cream and sage tones, keep the jar and its label exactly consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 1:1 · 1080p · 5s · amplitude: small. Best for: beauty and skincare hero clips placing a real product in a lifestyle setting.

Outfit & fashion try-on

Reference the model as one image and the garment as another, then describe the movement that shows the fit. Vidu keeps the person's face and the item true while animating the walk or turn. This is Vidu's answer to virtual try-on. For stylized versions, see the Vidu anime prompts.

17. Model spins in the dress

Using the reference images, the model from image 1 wears the dress from image 2 and does a slow graceful spin, the fabric flaring out with the motion, slow orbit matching her turn, soft editorial studio light against a seamless backdrop, elegant fashion mood, keep the model's face and the dress exactly consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 9:16 · 1080p · 6s · amplitude: medium. Why it works: one spin plus a matching orbit shows the garment from every side while two locked references keep the face and dress true.

18. Character tries on the sunglasses

Using the reference images, the character from image 1 raises the sunglasses from image 2 and slips them on, then tilts their head with a confident smile, slow push-in on the face, bright clean fashion light, simple bright background, keep the character's face and the exact shape and color of the sunglasses consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 9:16 · 1080p · 5s · amplitude: small. Best for: eyewear and accessory try-on clips that keep both the face and the product accurate.

19. Runner laces the branded shoe

Using the reference images, the athlete from image 1 sits on a bench and laces up the running shoe from image 2, then stands and steps forward ready to run, tracking macro on the shoe rising to a mid shot, crisp morning stadium light, energetic sportswear mood, keep the athlete's face and the shoe's design exactly consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 9:16 · 1080p · 6s · amplitude: medium. Why it works: the lace-and-stand action ties a real person to a real product in one continuous move without either reference drifting.

20. Model layers the coat over the outfit

Using the reference images, the model from image 1 pulls on the coat from image 2 over their outfit, adjusts the collar and turns to face the camera, tracking shot easing back with the movement, soft overcast street-style light, cool editorial mood, keep the model's face and the coat exactly consistent with their reference images.

Settings: Vidu Q2 Reference-to-Video · 2 reference images · 9:16 · 1080p · 6s · amplitude: medium. Best for: outerwear try-on and layering clips that keep the garment's cut and color true.

Scene & style transfer

Reference the subject and let Vidu re-render its look, lighting or setting. Point one reference at the content and describe the new style or environment, then keep the composition. This turns a still into a restyled moving shot. The cheat sheet lists the style keywords Vidu reads most reliably.

21. Character rendered in watercolor

Using the reference image, re-render the character in a soft loose watercolor illustration style with visible paper texture and gentle bleeding edges, they blink and smile slightly as a light breeze stirs their hair, slow subtle push-in, muted warm palette, keep the character's pose, features and composition consistent with the reference image.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 1:1 · 1080p · 5s · amplitude: small. Why it works: naming the medium plus a slow push-in restyles the reference into painterly motion without collapsing it back to photoreal.

22. Scene relit for golden hour

Using the reference image, relight the scene from the reference for warm golden-hour sunlight streaming in from the left, long soft shadows and a hazy glow, dust drifting in the light, slow gentle push-in, keep the layout, subjects and composition of the scene exactly consistent with the reference image, warm cinematic mood.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: small. Best for: changing the time of day on a locked scene while keeping every element in place.

23. Photo restyled as anime

Using the reference image, re-render the scene in a clean cel-shaded anime style with bold outlines and flat vibrant shading, the subject blinking and their hair swaying gently, a slow push-in adding life, warm nostalgic color, keep the composition, poses and framing consistent with the reference image, detailed anime sky.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 5s · amplitude: small. Why it works: Vidu reads style keywords cleanly, so naming cel-shaded anime plus one small motion transfers the look while the layout holds.

24. Room transformed to winter

Using the reference image, transform the room from the reference into a cozy winter version, frost on the windows and warm string lights glowing, soft snow visible falling outside, a slow drifting push-in through the space, warm amber interior light, keep the furniture, layout and composition consistent with the reference image, quiet homely mood.

Settings: Vidu Q2 Reference-to-Video · 1 reference image · 16:9 · 1080p · 6s · amplitude: small. Best for: seasonal and holiday variants of a real interior with the room kept recognizable.