Text to video
Describe subjects, interaction, camera language, lighting, shot order, dialogue, effects, and ambience in one production brief.
ByteDance video model · Available on Nano Banana Pro
Seedance 2.0 AI Video Generator
Turn a scene description or approved reference images into a short AI video with an optional generated soundtrack and platform controls.
01
LIVE Create a video with Seedance 2.0
Choose text, one image, or first and last frames, then set duration, resolution, aspect ratio, and generated audio before reviewing the point estimate.
Results
Generated content will appear here
02 / Overview
Seedance 2.0 at a glance
Facts reviewed:
- Developer
- ByteDance Seed
- Platform route
- Seedance 2.0
- Platform inputs
- Text and frame images
- Output
- Video with optional audio
- Platform duration
- 2–15 seconds
- Platform resolution
- 480p, 720p, or 1080p
Model brief
What is Seedance 2.0?
Seedance 2.0 is a video creation model from ByteDance Seed. The official launch describes a unified multimodal audio-video architecture that can use text, images, audio, and video as references. It emphasizes complex motion, physical plausibility, reference consistency, editing, extension, multi-shot storytelling, and synchronized stereo audio.
The full model can accept richer mixed references, but Nano Banana Pro currently exposes a narrower live workflow: text-to-video, one-image animation, or a platform first-and-last-frame route. This page does not offer audio or video reference uploads. Its 2–15 second duration range, 480p–1080p resolutions, aspect ratios, and point estimates are platform settings.
Use the standard route when a direction has moved beyond rough exploration and needs a deliberate production pass. Build a continuity brief before generating: lock wardrobe, prop dimensions, screen direction, eyelines, lens character, depth of field, motivated key light, color palette, weather, dialogue pronunciation, and ambient texture. Arrange these constraints around chronological story beats so visual and acoustic decisions reinforce the same scene.
Review a candidate at its intended playback speed, then inspect frame grabs at action peaks and shot boundaries. Check faces, hands, object edges, product markings, spatial geography, exposure shifts, and transition logic. Listen on headphones for clipping, channel imbalance, abrupt ambience, masked dialogue, and effects that arrive early or late. A polished-looking clip still needs this delivery review before approval.
For commercial handoff, package a provenance sheet with source licenses, performer consent, brand approvals, music clearance, generation timestamp, reviewer initials, version identifier, intended channels, distribution territory, and retention date. Mark unresolved defects and ownership conditions explicitly. This record keeps approved frames attached to their usage terms and gives legal, marketing, and production teams one auditable release reference.
03 / Capabilities 01 02 03 04 05 06
Key Seedance 2.0 capabilities
Use the model for motion-rich ideas while keeping official capabilities separate from the controls available in this integration.
Image to video
Animate one permitted still image while directing motion, performance, camera movement, timing, and sound.
Platform frame transition
Upload first and last images in the workspace to guide the beginning and destination of a continuous transition.
Joint audio-video generation
Enable generated audio and write dialogue, music, effects, and environmental sound alongside the visual timeline.
Complex motion
The official release highlights improved stability and physical plausibility for multi-subject interaction and demanding movement.
Multi-shot storytelling
Plan a compact sequence of connected shots; the official model supports high-quality audio-video output up to 15 seconds.
04 / Scene ideas
Popular Seedance 2.0 use cases
Use the model for motion-rich ideas while keeping official capabilities separate from the controls available in this integration.
Narrative previsualization
Explore staging, connected shots, performance, dialogue, and sound before committing to full production.
Action and movement concepts
Prototype sports, dance, interaction, and camera choreography that require coherent timing and physical motion.
Character-led social clips
Animate an approved character or portrait for short horizontal, square, or vertical campaign concepts.
Product and brand motion
Test controlled movement, lighting transitions, sound cues, and compact visual storytelling around approved assets.
Dialogue and performance scenes
Block short exchanges with explicit speaker order, camera timing, lip synchronization, room tone, and reaction beats.
Architecture and spatial transitions
Use compatible keyframes to plan camera travel through interiors, exteriors, changing light, and connected spaces.
05 / Prompt lab
Seedance 2.0 prompt examples
Each example loads the matching platform mode so you can revise the prompt in the live workspace.
SCENE 01text to video
Motion-rich sports sequence
A text-to-video brief with interacting subjects, camera continuity, physical detail, and synchronized sound.
View prompt
Cinematic indoor fencing practice between two experienced athletes. Begin with a low tracking shot beside their feet, rise into a medium two-shot as they exchange three controlled attacks and parries, then arc behind the winning athlete during the final touch. Accurate footwork, realistic blade flex, consistent uniforms and faces, cool overhead light, shoe squeaks, blade impacts, restrained room reverb, no titles or logos.
SCENE 02image to video
Character image performance
Animate an approved portrait while preserving appearance, clothing, and the original composition.
View prompt
Preserve the person’s facial identity, hairstyle, clothing, and background composition from the uploaded image. Add a natural breath, a brief glance toward the window, one small smile, and a slow camera push-in. Keep hands outside the frame, maintain realistic skin texture and lighting, soft city ambience, no dialogue, no added text.
SCENE 03start end to video
Designed opening-to-ending transition
Connect two compatible prepared frames with a clear transformation and camera path.
View prompt
Transition continuously from the uploaded opening frame to the uploaded ending frame. The camera moves forward through a curtain of drifting paper shapes as the lighting changes from cool dawn to warm sunset. Preserve the central subject and scene geometry, avoid cuts, keep movement smooth and physically coherent, add gentle paper rustle and distant wind.
SCENE 04text to video
Dialogue-driven micro drama
Coordinate restrained acting, shot progression, lip timing, and environmental sound in a compact scene.
View prompt
Naturalistic two-character scene in a late-night laundromat. 0–5s: locked medium two-shot while one character folds a blue shirt and says, “I kept your key.” 6–10s: slow push toward the second character as they answer, “Then you knew I would return.” 11–15s: hold eye contact while a washer stops behind them. Stable faces and wardrobe, accurate lip sync, fluorescent hum and machine click, no music or subtitles.
SCENE 05start end to video
Architectural first-to-last-frame move
Test a designed camera path and physically coherent lighting change between prepared architectural keyframes.
View prompt
Connect the uploaded daylight lobby frame to the uploaded evening rooftop frame in one continuous architectural move. Camera glides forward, rises through the central atrium, and clears the roofline as daylight shifts gradually to blue hour. Preserve structural lines, material colors, and travel direction; no cuts, no warped columns, subtle room tone changing into rooftop wind.
06 / Workflow
How to use Seedance 2.0
- 01
Choose an input mode
Start from text, upload one still image, or provide two compatible frames for the platform transition workflow.
- 02
Write the timeline
Describe the subject, action, shot order, camera movement, lighting changes, and ending in chronological order.
- 03
Coordinate audio
If generated audio is enabled, place dialogue, effects, music, and ambience beside the visual event they should match.
- 04
Review platform settings
Choose duration, resolution, and aspect ratio, then inspect the current point estimate before submitting.
Limits, pricing, and responsible use
Seedance 2.0 can still produce identity drift, unstable hands or objects, unintended cuts, inaccurate dialogue, or imperfect sound timing. Review the complete video and audio track before publication or client delivery.
ByteDance describes broader multimodal reference and editing capabilities than this page currently exposes. Nano Banana Pro accepts text and up to two frame images here; it does not accept audio or video reference files. The available 2–15 second range, resolutions, aspect ratios, generated-audio switch, and points belong to this integration rather than the official model product.
Only upload material you are authorized to use. Do not impersonate people, mislead viewers, violate privacy, infringe intellectual property, or remove required provenance and disclosure information from generated content.
Sources and fact review
Model capabilities and availability can change. Official ByteDance Seed sources and Nano Banana Pro platform settings are listed separately below.
07 / FAQ
Seedance 2.0 FAQ
Can I use Seedance 2.0 online on this page?
Yes. This page opens the Nano Banana Pro video workspace with Seedance 2.0 selected by default, while keeping the model selector available. An account and sufficient platform points are required to submit a task.
Is this Seedance 2.0 or Seedance 2.5?
This page and its live workspace use Seedance 2.0. Seedance 2.5 is a later model and is not represented by this integration.
Can Seedance 2.0 generate video from an image?
Yes. Choose image-to-video, upload one image you have permission to use, and describe the subject, camera, environment, pacing, and sound changes.
Does this page accept audio or video references?
No. The official Seedance 2.0 model supports broader mixed references, but this Nano Banana Pro workspace currently accepts text and frame images only.
Can Seedance 2.0 generate sound?
The official model supports joint audio-video generation. Enable generated audio in this workspace and describe dialogue, effects, music, and ambience in the prompt.
What duration and resolution are available here?
Nano Banana Pro currently offers 2–15 seconds with 480p, 720p, or 1080p output. The point estimate changes with the selected settings.
What is the first-and-last-frame option?
It is a Nano Banana Pro workflow that sends two compatible frame images to guide a continuous opening-to-ending transition. It is narrower than the official model’s full multimodal reference system.
Is Nano Banana Pro an official ByteDance product?
No. Nano Banana Pro is an independent multi-model platform. Seedance and ByteDance names and trademarks belong to their respective owners.
How do I reduce character drift in a multi-shot prompt?
Repeat a compact continuity block for identity, wardrobe, props, screen direction, and lighting. Change only the action and camera instruction for each timed shot, then inspect every transition before publishing.
What should I include when prompting dialogue?
Name the speaker, exact line, emotion, framing, and time window. Keep dialogue short, place sound cues beside the matching action, and ask for stable faces and accurate lip synchronization.
Related video models and tools
Render your next scene
Open Seedance 2.0 workspaceCreate your first Seedance 2.0 video
Return to the live workspace, choose the right input mode, and review the current point estimate before generating.