Skip to main content
ByteDance Seedance 2.5 generates coherent continuous videos or multi-shot sequences with native audio, strong prompt and layout adherence, first/last-frame control, and up to 50 multimodal references.

​
Text → Video

Generate one continuous take or a prompted multi-shot sequence with synchronized audio in a single pass. Select it with "mode": "text_to_video", or leave mode out and the inputs decide. Required: prompt, duration_seconds
Optional: aspect_ratio, generate_audio Resolution tiers: 480p, 720p, 1080p (send as tier).

​
Image → Video

Animate a required opening image, with an optional closing image. Output framing follows the input images. Select it with "mode": "image_to_video", or leave mode out and the inputs decide. Required: prompt, first_frame, duration_seconds
Optional: last_frame, generate_audio Resolution tiers: 480p, 720p, 1080p (send as tier).

​
Reference → Video

Guide characters, environments, style, motion, and sound with up to 30 images, 10 videos, and 10 audio clips (50 files total). Reference video and audio duration totals are each capped at 30.2 seconds. Select it with "mode": "reference_to_video", or leave mode out and the inputs decide. Required: prompt, duration_seconds
Optional: reference_images, reference_videos, reference_audios, aspect_ratio, generate_audio Resolution tiers: 480p, 720p, 1080p (send as tier).

​
Edit video

Change the clip you attach: add, remove or replace something, or restyle it. The result keeps the clip’s length and framing, and is billed on the clip’s length. Select it with "mode": "video_edit", or leave mode out and the inputs decide. Required: prompt, source_video
Optional: reference_images, reference_audios, generate_audio Resolution tiers: 480p, 720p, 1080p (send as tier).
Reference inputs take asset ids from POST /v1/uploads or an earlier output’s asset_id. Prices and allowed values come from the live catalog; GET /v1/models/seedance-2-5-fal is always current.