Models

Available nowIntegration verified 2026-08-11

Longer controlled video creation

Seedance 2.5 AI Video Generator

Create longer video sequences from text, an image, an existing clip, or up to 30 visual references with optional synchronized audio.

Try Seedance 2.5Explore capabilities

DreamMotion AI provides an independent workspace for this model. Current controls and credits reflect the verified integration.

4 workflows7 ratios30s max20–324 credits

Current integration

Seedance 2.5 at a glance

The Seedance 2.5 AI Video Generator supports text-to-video, image-to-video, video-to-video, and multi-reference workflows. Choose a supported duration, aspect ratio, resolution, and optional audio, then review the exact credit total before submitting the task.

Inputs

Ready

Text, 1–30 images, Video, 2–30 references

Resolution

Ready

480p, 720p

Aspect ratios

Ready

16:9, 9:16, 1:1, 4:3, 3:4, 21:9, Adaptive

Duration

Ready

4s, 5s, 6s, 7s, 8s, 9s, 10s, 11s, 12s, 13s, 14s, 15s, 16s, 17s, 18s, 19s, 20s, 21s, 22s, 23s, 24s, 25s, 26s, 27s, 28s, 29s, 30s, Auto

Current cost: 20–324 creditsProvider: APIMartView official model announcement

Create on this page

Try Seedance 2.5 in DreamMotion AI

The workspace is preselected for this model. Add a prompt or references, choose supported settings, and review the exact credit total before generating.

Open full workspace

Prompt framework

Build a controllable Seedance 2.5 request

Build stronger Seedance 2.5 prompts by defining the subject, ordered action, environment, camera movement, shot sequence, continuity, lighting, audio, and final output constraints.

Subject → Action → Environment → Camera movement → Shot sequence → Lighting → Audio → Output constraints
1

Plan the complete sequence

Describe the opening frame, ordered action, camera progression, continuity requirements, and final frame so a longer request has a clear visual arc.

2

Protect defining details

State which facial features, costume elements, product geometry, props, locations, and lighting relationships must remain stable as the shot develops.

3

Direct sound with picture

Place dialogue, ambience, music, and effects at specific visual beats so native audio supports the action and pacing instead of behaving like a separate layer.

Reusable example

Replace the bracketed details with your brief

Create a [4–30 second] [aspect ratio] video featuring [subject] completing [ordered action and story progression] in [environment]. Begin with [opening frame], move through [camera path and connected beats], and finish on [defined final frame]. Preserve [identity, costume, product, prop, location, or reference details] throughout the sequence. Use [lighting and color direction], place [dialogue, ambience, music, and sound effects] at specific visual moments, avoid [continuity errors or unwanted artifacts], and deliver at [480p or 720p] with coherent motion and synchronized audio.

Model overview

What is the Seedance 2.5 AI Video Generator?

Seedance 2.5 is a multimodal video creation model for generating a new shot from text, animating a still image, transforming existing footage, or combining several visual references. These four starting points let you choose between open-ended creation and tighter control over identity, composition, movement, environment, and style.

Seedance 2.5 is available in the DreamMotion AI video workspace for text-to-video, image-to-video, video-to-video, and references-to-video production. Choose a 4–30 second duration or automatic timing, one of seven aspect ratios, 480p or 720p output, MP4 or MOV format, and optional native audio while reviewing the exact credit requirement before generation. The same workspace keeps references, prompts, generation status, recent results, and failed-task refunds together, so the creative direction and production settings remain visible throughout the task.

How to use

How to create with Seedance 2.5

Follow the same four-step path from the first input to the final credit review, using only controls available in the current DreamMotion workspace.

  1. 1

    Choose a video workflow

    Start from text for a new scene, one image for controlled animation, an existing MP4 or MOV clip for transformation, or several images for reference-led direction.

  2. 2

    Write the sequence and reference roles

    Describe the subject, ordered action, environment, camera, continuity, lighting, sound, and final frame. Explain what each visual reference should control.

  3. 3

    Set duration and output

    Choose a 4–30 second duration, supported aspect ratio, 480p or 720p output, and whether the result should include synchronized native audio.

  4. 4

    Review credits and generate

    Confirm the displayed credit total and current balance before submitting. Confirmed failed tasks return their reserved credits automatically.

Workflow guide

Choose the right Seedance 2.5 workflow

Start from the asset that already contains the most important creative decisions. The right input mode reduces unnecessary prompt instructions and gives the model a clearer definition of what may change and what must stay consistent.

01Prompt-to-video creation

Text to Video

Start with a written brief when the scene does not depend on an existing asset. Describe the subject, ordered action, environment, camera path, lighting, pacing, and sound so the model can build the complete shot from one direction.

Best for: Original concepts, cinematic scenes, social clips, and early story development.

02Controlled still-image animation

Image to Video

Use one source image when composition, identity, product design, or art direction is already established. Prompt the motion that should change while identifying the facial features, objects, framing, and background details that need to remain stable.

Best for: Portrait animation, product reveals, campaign assets, and illustrated scenes.

03Existing-video transformation

Video to Video

Begin with a clip when timing and physical movement already exist. Direct the new visual treatment, materials, atmosphere, or camera character while protecting the original action arc and the moments that make the source footage readable.

Best for: Style transformation, shot refinement, visual experiments, and alternate treatments.

04Multi-reference video direction

References to Video

Combine several compatible references when different assets need to control character identity, wardrobe, environment, props, composition, or style. Give every reference a clear role and state which details must remain consistent throughout the sequence.

Best for: Character continuity, branded stories, product direction, and reference-led production.

Capabilities and best use cases

What Seedance 2.5 handles well

Explore Seedance 2.5 examples for complex motion, longer continuity, reference-led direction, and audio-aware storytelling, illustrated with official product media.

01

Full-body movement

Keep difficult physical action readable

The climbing sequence keeps the athlete, hold positions, body mechanics, and wall geometry readable while the camera follows a fast series of reaches and weight shifts.

Best for

Sports actionAthletic movementPerformance video
Recommended input
Text, image, or compatible motion references
Prompt focus
Body mechanics, action order, framing, camera response, and final pose

Capability examples use official dreamina seedance 2.5 product media from the cited official source.

Open official source

Model comparison

Compare video generation models

Choose by generation workflow, duration, creative control, output quality, and current credit range.

Kling 3.0Available

Text and image to video

11–322 credits
Seedance 2.0Available

Cinematic prompt-to-video shots

17–542 credits
Seedance 2.5Available

Longer cinematic sequences

20–324 credits
MiniMax H3Available

2K cinematic video

15–86 credits
Veo 3.1Available

Cinematic text-to-video scenes

5–5 credits
LTX-2.5Integration pending

Connected multishot story planning

Pricing pending

Choose Seedance 2.5 when the brief needs a longer 4–30 second production window, up to 30 visual references, MP4 or MOV source footage, and deliberate continuity across a more developed sequence. Choose Seedance 2.0 when the priority is output up to 4K, a shorter 5–15 second result, or a lower starting credit range. Both models support text, image, video, and multi-reference workflows with optional native audio inside DreamMotion AI.

Good to know

FAQ About Seedance 2.5

Clear answers about the current integration, credits, references, and generation behavior.

What can I create with the Seedance 2.5 AI Video Generator?+

DreamMotion AI supports Text to Video, Image to Video, Video to Video, and References to Video. Use text for a new scene, one image for controlled animation, an existing clip for transformation, or several visual references when identity, styling, environment, or composition should guide the result.

Which Seedance 2.5 video settings are available?+

The current workspace offers continuous 4–30 second durations plus automatic timing; 480p or 720p output; MP4 or MOV format; and landscape, portrait, square, standard, ultrawide, and adaptive aspect ratios. These controls reflect the verified APIMart integration used for submission.

How many reference images can I use?+

Image to Video uses one source image. References to Video accepts two to 30 images. Use compatible assets and assign each one a clear role such as identity, costume, product detail, environment, composition, prop, or visual style.

Can Seedance 2.5 transform an existing video?+

Yes. Video to Video accepts one MP4 or MOV reference clip between 2 and 30 seconds, from 480p to 720p, and up to 200MB. Describe what should change and explicitly protect the timing, action, identity, camera path, or scene details that should remain stable.

Does Seedance 2.5 support native audio?+

Yes. Enable native audio and describe dialogue, ambience, music, and sound effects at the visual beats where they should occur. The setting and exact credit requirement remain visible before submission.

How are Seedance 2.5 credits calculated?+

The total changes with resolution and duration. DreamMotion AI calculates the exact requirement from the selected settings and shows it beside the generation button. If the balance is too low, the button remains unavailable until enough credits are added.

What happens if Seedance 2.5 generation fails?+

DreamMotion AI returns the reserved credits after the task reaches a confirmed failed state.

Create with Seedance 2.5 today

Start inside the embedded workspace or open the full creator to keep more room for references, settings, and generated results.

Try Seedance 2.5Compare current models

10 free credits on sign-up · No credit card required