Seedance 2.5
Generate high-quality, audio-synced videos up to 30 seconds from text, images, video clips, and audio with ByteDance’s Seedance-2.5 multimodal model.
No workflow setup required · Fully customizable in AI-Flow
What This App Does
Seedance 2.5 is ByteDance’s flagship multimodal video model designed for production-ready, coherent video generation with native audio in a single pass. It supports mixed inputs (text, image, video, audio) and larger reference sets, enabling character consistency, motion/style transfer, and precise instruction following.
Key capabilities
Text-to-video with synchronized audio (dialogue, SFX, music)
Image-to-video from a first frame, with optional last frame control
Reference-guided generation using images, videos, and audios for style, motion, and timing
Video editing and extension while preserving camera motion and scene continuity
Multilingual prompts and strong adherence to complex, multi-subject directions
Native 30-second clip generation in one pass
Core inputs
prompt (string, optional) — Natural language description (up to 2000 characters). Optional if media inputs are provided.
image (string, optional) — First-frame image for image-to-video. Cannot be combined with reference_images/videos/audios.
last_frame_image (string, optional) — Requires image. Cannot be combined with reference sets.
reference_images (array, optional) — Up to 30 images for character/style/scene guidance. Reference as [Image1], [Image2], etc.
reference_videos (array, optional) — Up to 10 clips (max combined 30s) for motion/style/editing/extension. Reference as [Video1], [Video2], etc.
reference_audios (array, optional) — Up to 10 files (max combined 30s) for audio-driven generation and lip-sync. Requires at least one reference image or video. Reference as [Audio1], [Audio2], etc.
duration (integer, optional) — 4–30 seconds, or -1 for intelligent duration (required for editing mode).
aspect_ratio (string, optional) — e.g., 16:9, 9:16, 1:1, or adaptive (required for first/last-frame, editing, and extension).
resolution (string, optional) — 480p or 720p.
generate_audio (boolean, optional) — Toggle native audio (dialogue in double quotes), SFX, and music.
watermark (boolean, optional) — Add a watermark.
output_format (string, optional) — mp4 or mov.
seed (integer, optional) — Sets randomness; identical outputs not guaranteed.
Output - A URI to the generated video file.
Best practices
Put spoken dialogue in double quotes to enable lip-sync: "Remember this moment."
Be specific about subjects, actions, camera moves, lighting, and mood.
Label references in your prompt to combine modalities clearly: "The character from [Image1] performs the dance from [Video1] to the rhythm of [Audio1]."
Start with short durations (e.g., 5s) to iterate quickly, then scale up.
Notes and constraints
Do not combine first/last-frame inputs with reference_images/videos/audios in the same request.
reference_audios requires at least one reference image or video.
Use aspect_ratio = adaptive for first/last-frame, editing, and extension modes.
Intelligent duration (-1) is required for editing mode.
Example Input: seed=42, prompt="a golden retriever puppy running across a green meadow toward the camera, slow motion, warm afternoon light", duration=5, resolution=720p Output: A URI linking to the generated MP4 video.
How to Use the App
Here’s what you get when you open the app: a short form, one run button, your results — no nodes, no setup.
Seedance 2.5
App previewFill in the form
Aspect Ratio
16:9
Duration
Generate Audio
On by default
Image
Add images or files
Last Frame Image
Add images or files
+7 more options in the app
≈ 30–2096 credits / run
Get your result
Real output generated with this app
Under the Hood
Want more control?
This app is powered by a fully editable AI-Flow workflow. Change the model, prompts, layout, resolution — or connect it to other steps.
Image
Prompt
Slow, steady camera push-in toward the espresso cup, ending slightly closer with the cup rim still tack-sharp. Fine wisps of steam rise and curl from the coffee surface, drifting slowly to the right. The warm candle highlight in the background flickers gent...
Seedance-2.5
How to Use This Template
Step 1: Upload your file
In the 'Image' node, upload the file you want to process.
Step 2: Enter your text in 'Prompt' Node
Fill the 'Prompt' node with the required text.
No example available.
Step 3: Run the Flow
Click the 'Run' button to execute the flow and get the final output.
Customize the underlying workflow
Open the full workflow in the AI-Flow editor to swap models, rewrite prompts, or connect it to other steps.
Who is this for?
Perfect for professionals and creators looking to streamline their workflow
Video creators and filmmakers
Produce short-form, story-driven clips with consistent characters, controlled camera movement, and synchronized audio without complex pipelines.
Marketing and brand teams
Create product showcases, promos, and social content using reference images/videos for brand consistency and music- or VO-synced timing.
Content creators and social media managers
Rapidly generate meme-ready, vertical or horizontal videos with dialogue, SFX, and music in one pass.
Designers and art directors
Translate mood boards into moving visuals by combining style references, motion clips, and audio cues for precise creative direction.
Developers and prototypers
Integrate a reliable text/image/video/audio-to-video engine into apps and pipelines with clear I/O, aspect ratio control, and seed options.
Ready to create?
Start using this app
No workflow setup required — open the app and start creating in minutes
You Might Also Like
Explore other powerful templates to enhance your AI workflow
MiniMax H3 References to Video
Generate 2K videos from multimodal references with MiniMax H3. Guide subject, style, motion, and audio using up to 9 images, 3 video clips, and 3 audio clips cited by order in your prompt.
MiniMax H3 - First to Last Frame
Animate a first frame into a coherent 2K video (5–15s) with optional last frame matching, prompt-directed motion, and native stereo audio using MiniMax H3.
Gemini Omni Flash
Create cinematic AI videos from text, images, or existing footage with Gemini Omni Flash—multimodal text-to-video, image-to-video, and video-to-video with native sound design and coherent, real‑world motion.
Seedance-2.0-mini
Lower-cost, high-volume text-to-video and image-to-video generation with multimodal references and native, synchronized audio at up to 720p.
Shot to Motion
Turn a single photo into a short, cinematic motion clip. Upload a start frame, describe what happens next, optionally add reference images, and let the LLM build a precise prompt and consistent next frame before rendering a smooth video.
Storyboard to Cinematic Video with Seedance
Turn a 4-panel storyboard into a polished, multi-shot cinematic video. Generate the storyboard from a single prompt, then let its shot sizes, framing and camera rhythm drive one continuous 5-second spot with synchronized audio.
Frequently Asked Questions
What makes Seedance-2.5 different from typical text-to-video models?
It’s truly multimodal with native audio generation in the same pass as video, supports large reference sets (up to 30 images, 10 videos, 10 audios), and can generate up to 30 seconds natively while following complex, multi-subject prompts.
How do I add dialogue or ensure lip-sync?
Put dialogue in double quotes in your prompt, for example: "Remember this moment." The model generates synchronized lip movements and voice. You can also provide reference audio for rhythm or voice cues.
Can I combine first/last-frame images with reference images or videos?
No. First/last-frame mode cannot be combined with reference_images, reference_videos, or reference_audios. Choose either first/last-frame control or reference-driven generation.
When should I use aspect_ratio = adaptive?
Use adaptive for first/last-frame, editing, and extension modes, or when you want the model to pick the best framing automatically based on your inputs.
What durations are supported, and what is intelligent duration?
You can set duration from 4 to 30 seconds. Setting duration to -1 enables intelligent duration, where the model chooses an appropriate length. Editing mode requires -1.
How do reference sets work in prompts?
Upload references and cite them in your prompt using bracketed labels like [Image1], [Video1], and [Audio1]. For example: "The character from [Image1] performs the motion from [Video1] to the rhythm of [Audio1]."
Is generation reproducible with a fixed seed?
A seed can improve reproducibility, but identical results are not guaranteed due to the model’s stochastic nature and multimodal inputs.
Which output formats and resolutions are available?
Choose mp4 or mov for output_format, and 480p or 720p for resolution. 720p is recommended for higher visual quality.
Can I edit or extend an existing video?
Yes. Provide a reference video and describe what to change (editing) or what should happen next (extension). Use aspect_ratio = adaptive and duration = -1 for editing.
Does the model support multiple languages in prompts?
Yes. Prompts can be written in multiple languages, and the model maintains strong instruction following across languages.
What is AI-FLOW and how can it help me?
AI-FLOW is an all-in-one AI platform that allows you to build, integrate, and automate AI-powered workflows using an intuitive drag-and-drop interface. Whether you're a beginner or an expert, you can leverage multiple AI models to create innovative solutions without any coding required.
Is there a free trial available?
Yes, AI-FLOW offers a free trial to get you started. After that, you can purchase credits as needed—no subscription or long-term commitment required.
Can I integrate my API keys from providers like OpenAI and Replicate with AI-FLOW Cloud Version ?
Yes, you can easily integrate your existing API keys with AI-FLOW. If specified, nodes related to the API key provided will use your API key, significantly reducing your platform credit usage.