API
Console
Products
Docs
Pricing
Resources
API
Endpoint:
/bytedance/seedance-2.5-global-token/multimodal-video
seedance-2.5-global/multimodal-video Token
Seedance 2.5 Standard (global) multimodal video. Token-based billing.
Token Billing
Native Audio
Native 480p/720p + Upscaled 1080p/2k/4k
Smart Duration/Ratio
Multimodal Reference
reference-to-video
72.9 w
Playground
Playground
API
Pricing
Input
*
API Key
Please select API Key
*
prompt
A cinematic sequence: a rainy night in Times Square, New York. Neo-noir palette, anamorphic look, restrained and character-driven rather than a city montage. Reference usage: the reference image locks the Times Square framing, billboard density, neon palette and wet-asphalt reflections; reference video 1 sets the crowd rhythm and camera drift; reference video 2 sets the interior time-lapse energy; use the reference audio as the score, mixed low beneath the rain. [00:00-00:07] Wide shot of the avenue under heavy rain, billboards smearing red and blue across wet asphalt, yellow cabs pushing through standing water, umbrellas crossing the foreground in silhouette. Slow lateral dolly. Ambience: rain, tyre spray, distant sirens. [00:07-00:13] A man in a dark coat stands still at the centre of the crossing while the crowd streams around him; the camera pushes in slowly, the background compressing into bokeh. Ambience: footsteps, muffled traffic. [00:13-00:19] Cut to a warm bar interior seen through a rain-streaked window: silhouettes moving, amber light spilling onto the sidewalk; the camera drifts laterally along the glass, raindrops in sharp foreground focus. [00:19-00:25] Back outside, he turns and walks away from camera into the neon haze; the camera holds, rain intensifying, his silhouette dissolving into the light. Fade toward black with clean centre space. Style: neo-noir cinematic, anamorphic flare, saturated neon against deep blue shadow, high specular reflections on wet ground, shallow depth of field, subtle film grain. Camera: lateral dolly, slow push-in, drift along glass, final locked hold. No whip pans, no fast zooms. Audio: rain and city foley over a low jazz bed, no dialogue. Avoid: readable brand logos or legible billboard text; subtitles and watermarks; distorted faces or hands; duplicated or clipping pedestrians; jump cuts or flicker.
*
resolution
720p
*
duration
5
imageUrls
Add Input Item
(Max 30 items)
videoUrls
Add Input Item
(Max 10 items)
audioUrls
Add Input Item
(Max 10 items)
generateAudio
watermark
ratio
Adaptive
realPersonMode
conversionSlots
all
returnLastFrame
bitrateMode
standard
seed
outputFormat
mp4
omniReferenceTaskType
auto
Reset
Run
Result
Preview
JSON