Sign up and generate for free
By continuing, I acknowledge the Privacy Policy and agree to the Terms of Use. I also confirm that I am at least 18 years old
With AudioBest price on market for generations on the best AI models
Text, images, video, and audio in — multi-shot HD video with synchronized sound out. Coming to SpaceMotion.
Seedance 2.0 comes from ByteDance's Seed research team — a native multimodal model that reads text, images, video, and audio as one instruction and renders picture and sound together. It's the engine behind ByteDance's own creative tools, and it's coming to SpaceMotion's AI Toolkit alongside Veo, Sora, Kling and Seedance 1.5 Pro.
Seedance 2.0 is reference-first, multi-shot video generation — up to 12 reference files across images, video, and audio, real cuts inside a single pass, physics-accurate motion, and native audio with lip sync.
Images, video, and audio in.
Feed Seedance 2.0 up to 9 reference images, 3 video clips, and 3 audio tracks, and address them right in the prompt — @Image1, @Video1. Your characters, footage, and soundtrack become the raw material of the generation.
Edited sequences, not clips.
Seedance 2.0 generates natural cuts and transitions inside a single pass — an establishing shot, a close-up, a reveal — with characters and lighting held consistent across every edit.
Weight, impact, and fabric.
Collisions land with weight, fabric tears and folds like fabric, and multi-subject action — sports, choreography, crowds — moves the way the real world does.
Dolly zooms to handheld.
Call the shots in plain language — tracking shots, dolly zooms, POV switches, handheld shake. Native audio with precise lip sync is generated alongside every frame.
Seedance 2.0 is ByteDance's flagship video generation model, released in early 2026. It accepts text, images, video clips, and audio as input and generates multi-shot HD video with native synchronized sound — with director-level camera control and physics-accurate motion.
Text prompts, up to 9 reference images, up to 3 video clips, and up to 3 audio tracks — 12 reference files in total. You address them directly in the prompt (@Image1, @Video1, @Audio1) to cast characters, reuse footage, or drive scenes with a soundtrack.
Clips run from 4 to 15 seconds — or "auto", where the model picks the length that fits the action — at up to 1080p HD, 24fps, in widescreen, vertical, square, and cinematic 21:9 aspect ratios.
Yes. Audio is generated jointly with the video by default — dialogue with precise lip sync, sound effects, and ambience. You can also supply your own audio clips as references and the visuals are generated to match.
Instead of one continuous take, Seedance 2.0 composes sequences with real cuts and transitions inside a single generation — keeping characters, lighting, and style consistent across every shot, like edited footage.
Seedance 1.5 Pro takes text or a single starting image; 2.0 adds multimodal references (images, video, and audio), multi-shot editing with real cuts, stronger physics, camera direction, and longer 4–15 second clips. 1.5 Pro remains the faster, more economical choice for single-shot sounded clips.
Not yet — Seedance 2.0 is coming to SpaceMotion. In the meantime you can generate with Seedance 1.5 Pro, which shares the family's native audio and lip sync in a fast, economical package.
ByteDance launched Seedance 2.0 in China in February 2026 through Jimeng and Doubao, with global API availability following in April 2026.
Until it lands, generate sounded video with Seedance 1.5 Pro — same family, available today.
How multimodal reference-to-video works when Seedance 2.0 lands in SpaceMotion's AI Toolkit.
Collect the images, clips, and audio that define your subject — up to 12 files. Or start from text alone.
Describe the shots, cuts, camera moves, and dialogue, referencing your files inline — @Image1 walks through @Video1's set.
Seedance 2.0 renders the multi-shot sequence with synchronized audio in one pass — typically in under two minutes.
Not sure which Seedance fits your projects? Compare Seedance 2.0 vs Seedance 1.5 Pro and Seedance 1.0 to see which ByteDance video model fits your workflow.
ByteDance's flagship generation. Feed it reference images, video, and audio clips and it builds multi-shot video around them — real cuts, director-level camera work, physics-accurate motion, 4–15 second clips.
The audio-video workhorse, available on SpaceMotion today. Text- or image-to-video with multilingual dialogue, lip sync, and score generated in the same pass — 4–12 seconds, up to 1080p, seven aspect ratios.
More on Seedance 1.5 ProThe June 2025 original that topped video-generation leaderboards at launch. Text- or image-to-video with multi-shot storytelling, up to 1080p — silent output, before native audio arrived in 1.5 Pro.
From directors cutting multi-shot sequences to brands anchoring campaigns on real assets, Seedance 2.0 is for creators who bring their own material.
Try Seedance 1.5 Pro Cast from reference images, cut between shots, and call camera moves — an edited sequence from one prompt.
Your actual product across a multi-shot spot — anchored by up to 9 reference images.
Campaign films built around approved assets and existing footage, not lookalikes.
Multi-subject choreography with real weight and impact — physics the camera can follow.
Supply a soundtrack as a reference and generate visuals that move to it.
Recurring characters and formats, kept consistent across every episode by your references.
Start building with Seedance 1.5 Pro today — your workflow carries straight over when 2.0 arrives.
Sign up and generate for free
By continuing, I acknowledge the Privacy Policy and agree to the Terms of Use. I also confirm that I am at least 18 years old
With AudioBest price on market for generations on the best AI models
