Create 4–30 second AI videos from text or images, guide scenes with multimodal references, and generate synchronized audio in one workflow.
Choose 480p or 720p and review the credit estimate before you generate.
Explore all video modelsModel
Seedance 2.5
A continuous cinematic tracking shot through a rain-soaked city, with synchronized footsteps and passing traffic.
Longer generation, reference-led direction, native sound, and clear 480p or 720p output controls.
4–30 seconds
Generation duration
Up to 50
Reference inputs
Native
Generated audio
480p / 720p
Output resolution
Capabilities shown here match Oakgen's current Fal-backed production integration.
Capabilities
Every capability below maps to the current Fal-backed production integration.
Create a continuous clip from 4 to 30 seconds, giving ads, product stories, and narrative beats enough time to land without stitching together multiple short generations.
Generate continuous videos up to 30 seconds.
One continuous generated scene
Oakgen currently supports durations from 4 to 30 seconds.
Guide a generation with images, video, and audio references. Oakgen supports up to 30 images, 10 videos, and 10 audio files within the provider limits.
Guide a generation with up to 50 multimodal references.
Visual, motion, and sound references
Oakgen accepts up to 30 images, 10 videos, and 10 audio files.
Specify framing, subject movement, pacing, and camera direction in the prompt, then reinforce the result with references when the scene needs tighter creative continuity.
Use prompt and reference direction to guide action, framing, and camera movement.
Prompt direction paired with motion
Available in text-led and reference-led generation.
Create the video with synchronized dialogue, ambience, footsteps, and effects in the same generation, or switch audio off when you plan to finish sound separately.
Generate synchronized sound with the video.
Best experienced with sound
Audio generation can be switched on or off before generation.
Audio description: Dialogue, ambience, footsteps, and effects follow the timing of the on-screen action.
Creation gallery
Explore how longer motion, references, and sound can shape campaign, narrative, and social work.
Narrative storytelling
Brand campaign
Develop an opening, product moment, and closing beat within one connected scene.
Use separate references to guide recurring characters through a shared sequence.
Turn a product image and clear motion brief into campaign-ready footage.
Social video
Workflow preview
A simple three-step workspace for prompting, adding references, and generating.
Creation path
Video models
24 availableSeedance 2.5
Long scenes · multimodal
Seedance 2.0
Previous generation
Kling 3.0 Omni
Reference control
Selected model
Start with the model, then build the scene around the references and motion language that matter.
Model view
A clear view of the controls supported by the current production integration.
Oakgen currently supports durations from 4 to 30 seconds.
Oakgen accepts up to 30 images, 10 videos, and 10 audio files.
Audio generation can be switched on or off before generation.
These are the resolutions currently available through Oakgen's Fal integration.
Available now
Open the generator, choose text or image input, and create with the production model today.
Learn Seedance 2.5
Practical guides for prompts, reference planning, ads, and production workflows.
Review capabilities, production uses, reference inputs, and workflow advice.
Structure longer scenes with timed beats, camera direction, references, and audio.
Compare the two Oakgen generation workflows for your project.
Plan a hook, proof sequence, product moment, audio map, and final beat.
Build a story, continuity bible, reference pack, shot prompts, and review workflow.
Adapt practical prompts for beauty, sports, food, technology, fashion, and UGC ads.
Plan vertical hooks, safe framing, proof-led shots, captions, and variants.
FAQ
Seedance 2.5 is a ByteDance AI video model for text-led, image-led, and multimodal reference-led video generation with optional native audio.
Yes. You can use Seedance 2.5 in Oakgen's video generator for text-to-video and image-to-video generation.
Oakgen supports whole-second durations from 4 to 30 seconds for the current Seedance 2.5 integration.
Reference-led generation accepts images, videos, and audio. Current limits are 30 images, 10 videos, and 10 audio files, with up to 50 inputs in total.
Yes. Native audio generation is enabled by default and can be switched off before submitting a generation.
Yes. Choose the Seedance 2.5 image-to-video mode, upload a start image, write the motion prompt, and optionally provide an end image.
The current Oakgen integration supports 480p and 720p output. The selected resolution affects generation cost.
Text-led generation supports the aspect-ratio choices shown in the generator. Image-led generation follows the uploaded image automatically.
No. The provider requires at least one image or video reference when reference audio is supplied.
Open the text-to-video or image-to-video Seedance 2.5 generator, add your inputs, review the live credit estimate, and submit the job.
Available now
Start with a prompt or animate an image with native audio and up to 30 seconds of generated video.
Open Seedance 2.5