Cast it once, shoot every scene with Vidu Q4
Shengshu’s Vidu Q4 builds a scene from what you hand it — up to twelve reference images, tagged by name in the prompt, keep your people, products and places consistent — and returns up to 16 seconds at up to 4K, with dialogue and sound effects in the same pass. Or give it one photo as the first frame and describe what moves.
Runs on coins — the exact price shows before you confirm, and failed generations are refunded automatically.
- Max output
- 4K
- Clip length
- 5–16s
- References
- Up to 12 images
- AI audio
- Native, optional
Several cuts, one clip
Write the beats in order and say where the cut falls. Vidu Q4 changes shot inside one clip — wide, tracking, close-up — with the same rider and the same rain from beat to beat. This one is ten seconds; a clip runs to sixteen.
Illustration01
Illustration02
Illustration03
Illustration04The two films are Vidu Q4 renders — 720p, 10 seconds, sound on — encoded for the web, never edited. The stills beside them are storyboard illustrations drawn with OpenAI’s GPT Image model, not frames rendered by Vidu Q4.
One cast, every scene
Tag your reference in the prompt and she stays herself from the kitchen to the platform: the same face, the same hair, the same cardigan.
IllustrationReference
IllustrationKitchen, morning
IllustrationCrossing, golden hour
IllustrationSofa, evening
IllustrationPlatform, windStart from one still
Hand it a single picture as the first frame and describe only what moves. The subject and the composition come from your image.
Illustration
Illustration
IllustrationWhat Vidu Q4 does best
A cast that stays itself
Attach up to twelve reference images — a presenter, a product, a location — and tag them in the prompt as [@reference_image_1], [@reference_image_2] and so on. Faces, logos and outfits hold from the first second to the last.
Cuts inside one clip
Write the beats in order and say where the cut falls: a wide, then a tracking shot, then a close-up on the line. One generation, several shots, no stitching afterwards.
Dialogue and sound, in sync
Switch on audio and put the line in quotes: speech and sound effects arrive with the picture, timed to what happens on screen — at no extra cost.
From 540p drafts to 4K masters
Try the idea at 540p or 720p, then render the keeper at 1080p, 2K or 4K from the same prompt. Five resolutions, priced before you confirm.
A photo becomes the first frame
Hand it a single still and describe only what moves. The subject and the composition come from your picture; the output follows its shape.
Made for ads and stories
Product spots with your bottle in every shot, a presenter who looks the same in every scene, character pieces with spoken lines — the jobs where consistency is the whole point.
From references to a finished clip in three steps
Add your references
Up to twelve images of your people, products and places — or one photo to start from, or none at all for a text-only scene.
Write the scene
Tag each reference by number, write the beats in order, put any line in quotes, and switch audio on if it should speak.
Pick length and resolution
Any whole second from 5 to 16, at 540p, 720p, 1080p, 2K or 4K. The coin price shows before you confirm, and failed generations are refunded automatically.
Six prompts to try in Vidu Q4
[@reference_image_1] walks along the beach at sunset holding [@reference_image_2], then turns to the camera and says: "It goes everywhere I go." Waves and wind under the line.
A 16-second product spot: [@reference_image_1] the perfume bottle on wet black stone, a slow orbit as water beads on the glass, cut to a close-up of the cap, cut to a wide as one spotlight sweeps across. Low electronic pulse.
From my photo: the courier pushes off as the light turns green, the camera tracking alongside in the rain, then a close-up as he grins over his shoulder. Traffic hiss and rain.
[@reference_image_1] is the chef, [@reference_image_2] the kitchen. 0–5s: she plates a dish under warm pass lights. 5–10s: cut to a close-up as she wipes the rim. 10–14s: she slides it forward and says: "Service."
A 9:16 morning routine: [@reference_image_1] opens the curtains, pours coffee, laughs at her phone on the sofa — three shots, warm window light, soft room tone.
A misty fjord at dawn, a small red boat drifting across glassy water as cloud lifts off the ridge; slow push-in, 12 seconds, ambient birds and lapping water only.
Explore the other models
Browse AI templates by category
Make something worth sharing.
From film-grade scenes to product ads — or you, starring in the next viral clip. Free to start, right in your browser.