Seedance 2.0

Seedance 2.0 is the newest version, and its defining advances are multi-shot editing (cutting between several shots inside one generation), native synchronized audio, stronger real-world physics, and clips up to 15 seconds. Seedance 1.5 Pro also generates native audio but focuses on strong single-shot clips at a lower cost, and Seedance 1.0 Pro is the cost-efficient earlier flagship that does not generate audio. Pick 2.0 when you want an edited, sound-on sequence from one prompt.

Yes. Seedance 2.0 generates synchronized audio (sound effects, ambient sound, and lip-synced speech) in the same pass as the video, and audio is on by default. The per-second cost is the same whether audio is generated or not, so there is no price penalty for sound-on output. You can disable it if you want a silent clip.

Seedance 2.0 generates at 480p, 720p, and 1080p, with 720p as the default balance of speed and quality. It supports six aspect ratios (21:9, 16:9, 4:3, 1:1, 3:4, and 9:16) plus an automatic mode that picks a ratio to fit the prompt or input image, covering ultrawide cinematic, standard landscape, square, and vertical social formats.

A single generation runs from 4 to 15 seconds, or you can set duration to automatic and let the model choose a length that fits the prompt. That 15-second ceiling is longer than the 12-second maximum on Seedance 1.5 Pro and 1.0 Pro, which gives a bit more room for a full multi-shot beat in one clip.

Multi-shot editing means a single prompt can direct more than one shot, so the model cuts between angles or scenes inside one generation rather than producing a single continuous take. In practice you can describe a wide establishing shot followed by a close-up, and Seedance 2.0 renders both as a coherent edited sequence, which is unusual among text-to-video models.

Yes. Seedance 2.0 supports image-to-video, animating a still image as the opening frame. You can also supply an ending frame so the clip transitions from the start image to the end image along a controlled path. Input images can be JPEG, PNG, or WebP up to 30 MB, which makes it useful for bringing product photography or brand stills into motion.

On fal, Seedance 2.0 is billed per second of generated video, roughly $0.30 per second at 720p, with the same rate whether audio is on or off. Cost scales with resolution and duration since pricing is token-based on the output dimensions. Per-second rates change over time, so check current fal pricing for exact figures. Inside Masonry, generation is metered through your workspace.

Yes. Improved real-world physics is one of the model's stated strengths, and it handles demanding motion like sports, dancing, collisions, and pouring liquids while keeping subjects coherent. That makes it well suited to food sizzle shots, beverage pours, fitness, and action-driven social ads where flickering or distorted motion would break the illusion.

Both duration and aspect ratio default to an automatic mode, where the model picks a clip length (up to 15 seconds) and a frame shape that best fit your prompt or input image. You can also set them explicitly, choosing any duration from 4 to 15 seconds and any of the six aspect ratios from 21:9 ultrawide to 9:16 vertical. Auto mode is convenient for quick drafts, while explicit settings give you control when a placement needs a specific format or runtime.

Yes. Its synchronized audio includes lip-synced speech, so when a character speaks the mouth movements track the generated dialogue rather than looking dubbed. This makes it viable for talking-character social content and spokesperson-style spots where spoken audio needs to line up with the visuals.

Yes, through the prompt. Seedance 2.0 offers director-level camera control, so describing pans, push-ins, tracking shots, orbits, and angle changes carries into the generated motion. Combined with multi-shot editing, this lets you direct both how each shot is framed and how the sequence cuts together from a single text description.

Seedance 2.0 is built for commercial creative work like ads, social content, and product video, and Masonry routes generation through fal under its API terms. As with any AI video tool, review the current provider terms for your specific use case, especially for regulated categories or content involving real people's likenesses.

Describe the scene, the shots and how they cut together, the motion, and the audio you want, since the model generates all of these natively. For multi-shot output, name each shot and its framing (for example a wide establishing shot then a close-up). Including audio intent (ambient, effects, or a line of dialogue) gives you a more complete, sound-on clip in a single generation.

Seedance 2.0 is an AI video generation model from ByteDance, available inside Masonry, the AI creative agent teams use to produce marketing, product, and brand videos.

Open a Masonry canvas, pick Seedance 2.0 from the model selector, and describe the video you need: a product shot, an ad creative, a social post. Masonry generates it, then you refine, edit, and combine Seedance 2.0 with other models in one workspace.

Yes, you can start generating videos with Seedance 2.0 on Masonry's free tier, then scale up with higher limits and priority processing as your team grows.

Describe the shots and how they cut together, the motion, and the audio in one prompt. Seedance 2.0 generates multi-shot sequences and synchronized sound natively, so naming each shot and the sound you want gets you an edited, sound-on result in a single pass.

Seedance 2.0 is built by ByteDance. Inside Masonry it runs alongside 50+ image and video models, so your team can pick the right one for each brief without switching tools.