ByteDance video modelThree ways to create

Seedance 2.5

Give your next scene a beginning, a movement, and somewhere to go. Turn a written idea into video, bring a still image to life, or continue an existing clip with Seedance 2.5 in AudioX.

4–30 second generations480p to 4K optionsOptional generated audio

3

Creation workflows

4–30s

Generation duration

Up to 4K

Output resolution

Direct the scene from the starting point you have

Start with a written idea

Describe the subject, setting, movement, and mood in Text to Video. Choose a format for the destination, from a landscape film frame to a portrait social clip.

Give a still image movement

Use an image as the first frame and explain how the scene should unfold. Add an optional last frame when you have a particular ending in mind.

Continue the action

Open Video Extend with an existing clip and describe what happens next. Select 4–30 seconds of new material to continue a movement, reveal, or story beat.

Set the delivery format

Choose 480p, 720p, 1080p, or 4K and a duration that suits the idea. The workbench displays the credit cost for the settings you select.

Add context with references

In Text to Video, use reference images, videos, or audio to guide your brief. Explain the role of each reference and describe the action you want to see.

Plan sound with the picture

Enable generated audio when your scene calls for sound. Include useful direction for dialogue, atmosphere, or effects in the prompt, then review the result together with the visuals.

A video model for ideas, images, and the next shot

Seedance 2.5 is ByteDance’s video generation model. AudioX brings its text, image, and extension workflows into the tools you already use.

A good video brief connects what is in the frame with what changes over time. Establish the subject and setting, describe the main action, and finish with the camera position or moment you want to reach. Seedance 2.5 gives that brief a generation window from 4 to 30 seconds.

Choose the starting point that gives you the most useful control. A prompt leaves room to invent the scene. A first-frame image anchors its initial appearance. An existing clip supplies the action to continue. For Image to Video, a last-frame image can also guide the destination.

The controls described here are the ones available in AudioX’s Seedance 2.5 tools. Set your duration, resolution, and audio option in the workbench before generating, and review its current credit estimate.

Explore motion, atmosphere, and visual worlds

Reference examples published on the Seedance 2.5 pages at Clipdance and invideo. These clips and stills were sourced from those sites; they were not generated in AudioX. Video previews play muted, and results depend on the source material and settings.

Animation: expressive close-ups and energetic motion. Source: Clipdance.
Cinematic action: atmosphere, framing, and changing viewpoints. Source: Clipdance.
Shared worlds: a scene with several people and layered spaces. Source: invideo.
A futuristic aircraft with bright blue engines above a cloud-covered landscape.
Science fiction: light and scale establish a visual direction. Source: Clipdance.
A lone warrior stands above the sea on a dramatic rocky cliff at sunset.
Fantasy: a wide environment sets the mood for a story. Source: Clipdance.
A red railway signal framed against the steel roof of a monochrome station.
Visual storytelling: a simple focal point in a detailed setting. Source: Clipdance.

Pick your starting point

All three Seedance 2.5 workflows offer resolution and duration controls. The right choice depends on what you already have and what the next generation needs to do.

WorkflowStart withUse it for
Text to VideoA prompt, with optional referencesDevelop an original scene, compare concepts, or describe a shot before you have a starting image.
Image to VideoA first frame and an optional last frameAnimate artwork, a product visual, or a character image while guiding the scene from a defined starting point.
Video ExtendAn existing video and a continuation promptAdd the next beat to a clip. The duration you select is the amount of new video to generate.

Match the brief to a clear creative decision

Keep the next generation focused enough to judge whether it works.

  • A product needs a reveal — Start with a clear product image. Describe a single camera move, the lighting, and which product details should stay recognizable.
  • A character needs a moment — Give the subject one readable action and an emotional direction. A glance, a turn, or a small interaction can tell more than several competing actions.
  • An existing clip ends too soon — Use Video Extend to continue from the scene already in motion. Explain the next event and how the camera should follow it.
  • A longer story needs several scenes — Plan the sequence as individual shots. Generate or extend the pieces you need, then review continuity and assemble them in your editing workflow.

From a rough idea to a reviewable scene

Four steps to create with Seedance 2.5 in AudioX.

  1. 1

    Choose the input

    Open Text to Video, Image to Video, or Video Extend. The links on this page select Seedance 2.5 for you.

  2. 2

    Describe what happens

    Write the setting, subject, action, and camera movement. Add the images, video, or reference material needed by your chosen workflow.

  3. 3

    Set the format and review credits

    Choose the duration and resolution, and decide whether to generate audio. Review the workbench’s current credit cost before submitting.

  4. 4

    Watch, refine, and continue

    Review the action, composition, and sound. Make a focused adjustment for another take, or bring a suitable result into Video Extend when the story needs more time.

Write a prompt with a visible sequence

Describe the decisions a viewer can actually see and hear.

1. Establish the first moment

Begin with the subject, location, and lighting: “A cyclist waits at a quiet street corner just after sunrise.” This gives the action a setting.

2. Give the subject one main action

Continue with a clear movement: “She looks over her shoulder, then rides toward the bridge.” Keep the action practical for the chosen duration.

3. Direct the camera separately

Describe framing and movement in their own sentence: “The camera tracks alongside at eye level, then eases into a wider view.”

4. Explain the job of each reference

Say which reference supplies the character, setting, motion, or sound. For image animation, focus on what should move and what should remain visually consistent.

5. Write the ending and the sound

Finish with a destination for the action. If audio is enabled, describe the intended atmosphere or dialogue. For an extension, connect the first new beat to the end of the uploaded clip.

Avoid packing many unrelated actions into a short clip or asking the camera to move in conflicting directions. Review each result and change the part of the brief that matters most.

Choose time and detail for the scene you need

Seedance 2.5 offers 480p, 720p, 1080p, and 4K output choices in AudioX. A lower resolution can help when you are comparing ideas; choose the output resolution that suits the version you intend to use. The workbench shows how the selected settings affect credits.

Text to Video offers six aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9. Plan for the destination before you generate, whether it is a vertical feed, a square placement, or a wide scene.

Choose 4–30 seconds for a new generation. In Video Extend, that number describes the new segment to add, not the total duration of the uploaded clip and its continuation.

Create with Seedance 2.5 in AudioX

Choose the tool that matches the material on your desk.

Text to Video

Build a scene from a written brief and optional image, video, or audio references.

Create from Text

Image to Video

Bring a first-frame image into motion, with an optional last frame to guide the finish.

Animate an Image

Video Extend

Continue an existing video with a prompt describing the next action, reveal, or ending.

Extend a Video

Seedance 2.5 questions

Practical details about the three AudioX workflows.

What is Seedance 2.5?

Seedance 2.5 is a ByteDance video generation model. In AudioX, you can use it to generate from text, animate an image, or extend an existing video, with selectable duration, resolution, and generated audio.

How long can a generation be?

Choose a duration from 4 to 30 seconds. In Text to Video and Image to Video, this is the length of the generated clip. In Video Extend, it is the length of the new continuation.

Which resolutions can I use?

The available output choices are 480p, 720p, 1080p, and 4K. Select the option in the workbench and check the current credit estimate before you generate.

Can I provide both a first and a last frame?

Yes. Image to Video uses your first-frame image and supports an optional last-frame image. Describe the movement or transition between the images so the generation has a clear direction.

Can I use references in Text to Video?

Yes. The Text to Video workflow supports image, video, and audio references. Use the input controls in the workbench and explain how each reference should influence the scene.

Does Seedance 2.5 generate audio?

Generated audio is optional. Enable it when you want sound with the video and include relevant audio direction in your prompt. Review the generated sound and its timing before using the result.

How does Video Extend work?

Upload an existing clip, describe what should happen next, and choose how many new seconds to generate. Use a prompt that continues the subject, setting, and movement already present at the end of the clip.

How many credits does Seedance 2.5 cost?

The workbench displays the current credit cost for your chosen workflow and settings before you submit. Check that estimate after changing duration, resolution, or other inputs. Visit AudioX’s pricing page for available credit plans.

Which aspect ratios are available for Text to Video?

Choose 16:9, 9:16, 1:1, 4:3, 3:4, or 21:9. Pick the shape for your intended placement before generating so you can compose the scene around it.

Were the examples on this page generated in AudioX?

No. The gallery contains reference clips and images published on Clipdance’s and invideo’s Seedance 2.5 pages. They are presented as visual inspiration with source labels. They do not guarantee a particular result from a AudioX generation.

Prepare the source. Review the scene.

Good inputs and a careful final review help you turn a promising take into usable work.

Before you generate

  • • Use images, video, and audio you have permission to work with.
  • • Write the intended action and the details to preserve.
  • • Check duration, resolution, audio, and the displayed credit cost.

Before you publish

  • • Review faces, text, product details, and motion across the clip.
  • • Listen to generated audio and check its timing.
  • • Confirm the result meets your project and platform requirements.

Make the next scene yours

Start with an idea, an image, or a clip worth continuing. Open Seedance 2.5 in AudioX and give your next shot a clear direction.

Try Seedance 2.5