Longtake

SEEDANCE 2.5 AVAILABLE NOW

50 references. One coherent film. Direct images, video and audio from a single prompt.

Seedance 2.5 · up to 30s · 4K output

NEW

SEEDANCE 2.5TECHNICAL BREAKDOWN

NOTHING DRIFTS

LightingCharacter identityScene
Portrait with a selected face region

REGION EDIT

00:12
00:08
CITY ATMOS · 00:36
VOICE LINE · 00:14
+38

50 REFERENCES

One generation, everything at once

30S

Six times the runway in one continuous take

00:0000:2000:30
NEW

TOTAL CONTROL OVER YOUR PRODUCTION WITH SEEDANCE 2.5

One continuous take

ONE SHOT. NO CUTS. NO COMPROMISES.

Hold the frame and let the moment unfold. No cuts, no transitions, no zooms — one continuous take with movement, timing and character continuity intact.

Capability 1
Seedance motion control

DRIVE THE MOTION WITH SEEDANCE

Point the camera path and the subject's move, then hand it to Seedance to execute frame by frame.

Capability 2
A cinematic long-form video scene with a three-minute duration marker
Long-Form Video

TAKE THE WHOLE STORY FROM START TO FINISH

Generate up to 30 seconds at once, extend existing clips, or create videos up to 180 seconds in Ultra-long mode.

Capability 3
A cinematic scene directed with timestamp-specific instructions
Timestamp Control

DIRECT EVERY SECOND OF THE STORY

Add timestamps to decide exactly when actions, dialogue, camera changes, and transitions happen.

Capability 4
Motion transferred from a human performer to a different subject
Creative Transfer

TRANSFER MOTION TO ANY SUBJECT

Copy the movement, camera motion, and rhythm of a reference video while changing its subject or setting.

Capability 5
A selected video region changed while the rest of the shot stays intact
Region Editing

EDIT WITHOUT REGENERATING

Replace objects, refine characters, or remove unwanted details while keeping the rest of the video unchanged.

Capability 6
The same cinematic scene rebuilt from a different camera viewpoint
Viewpoint Control

CHANGE THE CAMERA VIEWPOINT

Rebuild the same scene from a new angle while preserving the subject, action, and environment.

Capability 7
Multiple characters remaining visually consistent in the same scene
Multi-Person Reference

KEEP EVERY CHARACTER CONSISTENT

Assign a separate reference to each character so faces, clothing, and roles remain distinct across every shot.

Capability 8
A green-screen performance placed into a cinematic action environment
Green-Screen Reference

PLACE PERFORMANCES IN ANY WORLD

Preserve the actor’s movement and timing from green-screen footage while generating an entirely new environment.

Capability 9
A 3D blockout converted into a finished cinematic shot
White-Model Control

TURN BLOCKOUTS INTO FINAL SHOTS

Use a 3D blockout to control the layout, character positions, and camera movement of the final scene.

Capability 10
Two video clips connected through a generated transition
Video Transition

BRIDGE TWO CLIPS SEAMLESSLY

Provide the starting and ending clips, and let Seedance generate the missing transition between them.

Capability 11
Multiple storyboard frames combined into one connected video sequence
Multi-Grid Storyboard

TURN STORYBOARDS INTO SEQUENCES

Arrange key moments in a multi-grid storyboard, then generate them as one connected video sequence.

Capability 12

GOT ANY QUESTIONS LEFT?

We’ve answered the most frequently asked questions

What is Seedance 2.5?

ByteDance’s flagship AI video model. It generates cinematic clips up to 30 seconds in native 4K, with synchronized sound, up to 50 references, and region-level editing — all in a single pass.

What’s new compared to Seedance 2.0?

Seedance 2.5 expands creative control with longer native clips, synchronized sound, multi-material references, region editing, and more precise shot-level direction.

How long and what resolution can videos be?

A single generation can run for up to 30 seconds and deliver native 4K output, depending on the selected generation settings.

Does it generate sound?

Yes. Seedance 2.5 can generate synchronized dialogue, effects, and atmosphere alongside the picture in the same generation.

How many references can I use?

You can direct a generation with up to 50 image, video, and audio references in one prompt.

Can it keep the same character across scenes?

Yes. Character identity, lighting, props, and scene details can remain coherent across a multi-shot sequence when they are included as references.

How do I get the best results?

Start with a clear shot description, add focused visual or audio references, and use the dedicated controls for camera, lighting, pacing, and motion.

How do I access Seedance 2.5 on Longtake?

Describe the film you want to make, select Generate now, and continue in Longtake’s AI video generator.