Skip to main content
Upgrade
Loading account

Seedance 2

Seedance 2 creates short videos from text, images, video references, or first and last frames, with optional generated audio. Try Seedance 2 on Nuzza with duration, aspect-ratio, resolution, and bitrate controls in one queued workflow.

Describe what you want to create with Seedance 2

Loading inspiration gallery.

    Seedance 2: Direct a Complete Short Scene from a Prompt

    A Seedance 2 workflow combining a prompt, image references, and video references into one short cinematic clip

    Describe the subject, setting, action, camera path, pacing, lighting, and intended sound in one readable brief.

    A three-shot night greenhouse sequence planned from one prompt with matching action, camera direction, and generated sound cues

    Direct a Complete Short Scene from a Prompt

    Describe the subject, setting, action, camera path, pacing, lighting, and intended sound in one readable brief. Seedance 2 can turn that direction into a short audiovisual sequence, including multiple shots when the prompt makes their order clear. Keep the request focused enough to evaluate motion, continuity, and timing.

    Try For Free
    Start and end frames guiding a bicycle courier from a still street corner into a moving rainy-night shot

    Shape Motion Between a First and Last Frame

    Begin with one image to anchor the composition, or add an end frame to define where the clip should arrive. Explain what moves, what stays recognizable, and how the camera travels between those anchors. The frames guide the transition; they do not guarantee an exact path, so review intermediate motion and subject detail.

    Try For Free
    Subject, wardrobe, location, and motion references converging into one coherent Seedance 2 scene

    Coordinate Multiple Image and Video References

    Add up to nine images and three videos, then assign each source a clear job: subject identity, wardrobe, location, composition, movement, camera rhythm, or visual tone. A role-based prompt helps the model reconcile the inputs. Nuzza does not expose imported audio references for this workflow, even though the upstream model supports them.

    Try For Free
    One Seedance 2 clip arranged across landscape, square, portrait, and ultrawide output frames with audio and bitrate controls

    Match the Clip to Its Delivery Frame

    Choose 480p, 720p, 1080p, or 4K; use auto duration or select 4–15 seconds; and frame the result as ultrawide, landscape, classic, square, portrait, or vertical. Generated audio is on by default, while standard and high bitrate options let you balance the requested encode against file size.

    Try For Free

    How to Create a Video with Seedance 2

    01

    Choose the Workflow That Matches Your Source

    Use Text when the scene exists only as an idea. Use Image when one frame should anchor the shot, Reference when several images or videos should guide different attributes, and Frames when you need explicit starting and ending images. This choice determines which uploads and controls appear in the composer.

    02

    Write the Shot and Assign Every Reference

    Describe the subject, environment, action, camera movement, pacing, lighting, and sound in the order a viewer should experience them. For reference-led work, state what each image or video contributes and which details should remain stable. Avoid asking one short clip to resolve several unrelated actions or conflicting sources.

    03

    Set the Output, Queue the Request, and Review

    Choose resolution, auto or 4–15 seconds, aspect ratio, generated audio, and bitrate, then review the request-time credit quote and submit the queued job. Inspect the finished clip for identity drift, unintended transitions, timing, sound, and framing. Revise one variable at a time when the result needs another pass.

    Practical Ways to Use Seedance 2

    A three-shot product previsualization using the same unbranded travel speaker across a studio, train platform, and night landscape

    Test a Multi-Shot Product Story Before Production

    Combine a product image, a location reference, and a concise shot sequence to explore how a campaign idea moves before a full shoot. Define the opening composition, product action, camera transition, and final frame. Use the result to discuss pacing and direction, not as proof that logos, labels, or geometry will remain exact.

    Try For Free
    A reference camera arc and dancer pose translated into an original rooftop movement study at sunset

    Translate Reference Motion into a New Scene

    Use a reference video to communicate camera rhythm or body movement, then pair it with original subject and environment images. Tell the model which motion qualities to follow and which visual details belong to the new scene. This is useful for exploring direction while avoiding a claim that every beat or identity will transfer exactly.

    Try For Free
    Opening, development, and closing frames of a vertical café story with visualized ambience and dialogue cues

    Explore a Short Story with Generated Audio

    Write a compact opening, development, and closing beat for a vertical or landscape social concept, then leave generated audio on when sound effects, ambience, or speech help communicate the idea. Keep dialogue brief and review synchronization, pronunciation, continuity, and platform requirements before treating the clip as publishable.

    Try For Free

    How to Get More Reviewable Seedance 2 Results

    Shot structure
    Recommended Seedance 2 shot with one ceramic dog action and one gentle camera push

    Keep one action and one camera path readable

    Avoid example where unrelated actions and conflicting camera moves break one short shot

    Avoid stacking unrelated actions into one short clip

    Reference roles
    Recommended reference board where subject, wardrobe, location, and camera motion each have one clear role

    Give every image and video reference one clear job

    Avoid example where conflicting subjects, wardrobe, and motion references create a confused result

    Avoid mixing sources with conflicting subjects or movement

    Preservation
    Recommended start and end frames preserving a coral kettle while only grass, cloth, and steam begin to move

    Name the identity, composition, and frame details to preserve

    Avoid example where the kettle shape, handle, framing, table perspective, and lighting drift between frames

    Avoid assuming references guarantee exact continuity

    Seedance 2 vs Seedance 2 Fast vs Kling 3.0

    Best for

    Seedance 2

    High-resolution short clips that need broad visual references, start/end frames, or detailed output control.

    Seedance 2 Fast

    Faster same-family iteration when 480p or 720p is sufficient for review.

    Kling 3.0

    Shot-by-shot planning that benefits from separate prompt segments and shot-type control.

    Nuzza workflows

    Seedance 2

    Text to video, image to video, reference to video, and start-and-end-frame video.

    Seedance 2 Fast

    Text to video, image to video, reference to video, and start-and-end-frame video.

    Kling 3.0

    Text to video, image to video, and start-and-end-frame video.

    Reference inputs

    Seedance 2

    Up to nine images and three videos in the dedicated reference workflow.

    Seedance 2 Fast

    Up to nine images and three videos in the dedicated reference workflow.

    Kling 3.0

    No dedicated multi-reference video workflow in the current Nuzza mapping.

    Output and direction controls

    Seedance 2

    480p–4K, auto or 4–15 seconds, seven aspect-ratio options, generated audio, and standard or high bitrate.

    Seedance 2 Fast

    480p or 720p, auto or 4–15 seconds, the same seven ratio options, generated audio, and standard or high bitrate.

    Kling 3.0

    3–15 seconds, landscape, vertical, or square framing, native-audio toggle, multi-prompt, shot type, and guidance controls.

    Choose it when

    Seedance 2

    Reference breadth and output range matter more than choosing the speed-focused tier.

    Seedance 2 Fast

    You want the same Seedance workflow structure with a speed-focused, lower-resolution ceiling.

    Kling 3.0

    Explicit segmented shot planning matters more than Seedance 2's broader multi-reference input.

    Choose Seedance 2 when you need the widest output range in this set and want to coordinate several image or video references. Choose Seedance 2 Fast when the same four-workflow structure and reference limits matter but 720p is enough for faster iteration. Choose Kling 3.0 when separate prompt segments and shot-type controls are more useful than a dedicated multi-reference workflow. Nuzza keeps each option inside the same queued composer pattern, so compare the controls shown for the selected model before submitting.

    FAQs about Seedance 2

    What is Seedance 2?

    Seedance 2 is Nuzza's display name for ByteDance's Seedance 2.0 video model. It is designed for short multimodal video generation, including multi-shot clips with generated sound. The upstream model can interpret text, images, audio, and video, while Nuzza currently exposes text, image, video-reference, and start/end-frame workflows rather than imported audio references.

    What is the best AI video generator?

    The best AI video generator depends on the input, intended result, budget, and controls you need. Seedance 2's Nuzza workflow is a strong choice because it is built around a full-quality, reference-led Seedance workflow.

    Is Seedance 2 free to use?

    The page is free to browse. Running a task requires Credits, and the current cost is shown before you submit.

    What can I use Seedance 2 for?

    Seedance 2 is useful for previsualization, motion study, and social concept. Combine a product image, a location reference, and a concise shot sequence to explore how a campaign idea moves before a full shoot. Define the opening composition, product action, camera transition, and final frame. Use the result to discuss pacing and direction, not as proof that logos, labels, or geometry will remain exact.

    What inputs does Seedance 2 support?

    You can create from text, animate a starting image, guide a clip with multiple image and video references, or provide start and end frames. Every workflow uses a prompt. Choose the mode that matches the source material you actually have instead of adding references that do not contribute a clear subject, setting, motion, or composition role.

    Do I need to sign in to use Seedance 2?

    A signed-in account and sufficient Nuzza credits are required to submit a generation job.

    What should I check in the video result?

    Watch the complete result and check subject identity, motion, camera continuity, frame edges, text, and audio synchronization. Compare important details with the source or prompt before publishing.

    What length, resolution, and aspect ratios does Seedance 2 support?

    Nuzza exposes automatic duration or a selected length from 4 to 15 seconds. Resolution options are 480p, 720p, 1080p, and 4K. Aspect-ratio options are auto, 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16. Availability describes the requested settings, not a guarantee that every visual detail will render perfectly.

    Can I use Seedance 2 videos commercially?

    Nuzza does not claim ownership of an output merely because it was generated through the service, but that is not a promise of commercial clearance. You remain responsible for rights in prompts, references, likenesses, voices, trademarks, and the finished output. Applicable law and separate provider or platform terms may also affect a planned use, so review them for your project.