Bytedance family
# Seedance 2 Advanced multimodal video generation with native audio. ## Overview Seedance 2 is a state-of-the-art multimodal video generation model developed by ByteDance, designed to bridge the gap between static imagery and cinematic motion. Unlike traditional video models that treat audio as an afterthought, Seedance 2 features native audio synthesis, allowing for the simultaneous generation of synchronized dialogue, sound effects, and background music. This integration ensures that the temporal flow of the video is perfectly aligned with the auditory experience, creating a cohesive narrative output. Beyond its audio capabilities, Seedance 2 excels in character and style consistency. By utilizing a robust reference input system, users can provide multiple images, videos, or audio clips to guide the generation process. This makes it an ideal tool for professional creative workflows where maintaining a specific brand aesthetic or character appearance across multiple clips is essential. With intelligent duration control and flexible aspect ratio management, the model adapts to various production requirements, from short-form social media clips to long-form cinematic sequences. ## Use Cases * **Brand Character Consistency:** Maintain a consistent brand mascot or spokesperson across multiple video assets using the reference image input system. * **Social Media Ad Production:** Generate high-quality 9:16 or 9:21 vertical video content with native, synchronized audio for platforms like TikTok, Instagram Reels, and YouTube Shorts. * **Automated Storyboarding:** Transform static concept art into animated sequences with specific motion cues and background music to pitch creative concepts to clients. * **Localized Marketing Campaigns:** Use reference audio and lip-sync features to generate localized versions of video ads with consistent character movement and regionalized dialogue. * **Dynamic Product Showcases:** Create immersive product reveal videos by combining high-resolution product images with motion-transfer reference videos. * **Narrative Content Creation:** Develop short-form storytelling content where the model handles the complex synchronization of sound effects and dialogue based on text prompts. ## Parameters Seedance 2 offers a comprehensive suite of configuration options to fine-tune your output. Users can adjust parameters ranging from basic prompt engineering and resolution settings to complex multimodal reference inputs. Please refer to the Pixloop parameter configuration panel for specific input types, including seed control, reference image/video/audio uploads, and duration settings. ## Tips for Best Results * **Leverage Reference Inputs:** For maximum character consistency, provide at least 3-5 high-quality reference images and reference them explicitly in your prompt using the [Image1], [Image2] syntax. * **Use Intelligent Duration:** If you are unsure of the optimal length, set the duration to -1. This allows the model to analyze your prompt and determine the most natural pacing for the generated scene. * **Dialogue Precision:** When generating dialogue, wrap your spoken text in double quotes within the prompt to signal the model to prioritize lip-sync and voiceover generation. * **Audio-Driven Motion:** Use the reference_audios parameter to influence the rhythm of the video. The model will align motion peaks with the beats or speech patterns found in your uploaded audio files. * **Adaptive Aspect Ratios:** For multi-platform campaigns, use the 'adaptive' aspect ratio setting to allow the model to optimize the composition based on your input assets, reducing the need for manual cropping. ## About Seedance 2 is a flagship multimodal model from ByteDance, representing a significant leap in generative video technology. It builds upon the lineage of high-performance diffusion models, specifically optimized for the tight integration of visual and auditory data. By leveraging advanced architectural improvements in temporal consistency and cross-modal attention, Seedance 2 provides marketing teams with a powerful, production-ready tool for high-fidelity video synthesis.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| image | image | No | null | First frame image. Cannot be combined with reference images. |
| prompt | string | Yes | null | Text prompt for video generation. Recommended under 1000 words. |
| duration | integer | No | 5 | Video duration in seconds. Seedance 2.0 range [4,15], or -1 for intelligent duration. |
| resolution | select | No | "720p" | Video resolution. Seedance 2.0 (base) supports up to 4k. |
| aspect_ratio | select | No | "adaptive" | Video aspect ratio (maps to `ratio`). Default adaptive. |
| generate_audio | boolean | No | true | Generate synchronized audio (voice/SFX/music). Put dialogue in double quotes in the prompt. |
| last_frame_image | image | No | null | Last frame image. Only works with a first frame image. Cannot be combined with reference images. |
| reference_audios | audio | No | [] | Up to 3 reference audios (total <= 15s) for audio-driven/lip-sync. Requires at least one reference image or video. Reference as [Audio1] etc. |
| reference_images | image | No | [] | Up to 9 reference images for character consistency/style/scene. Reference as [Image1], [Image2] in the prompt. No real human faces. |
| reference_videos | video | No | [] | Up to 3 reference videos (total <= 15s) for motion transfer/style/editing. Reference as [Video1] etc. |