Ask anything...
Now on Venice

ByteDance

State-of-the-art cinematic video generation.

Start creating
ByteDance

Three ways in

Start from text, an image, or your own references

Generate up to 30 seconds of cinematic video with native audio from text, an image, or up to 50 references. Available on Venice.

Text to Video

Type a prompt, get a cinematic clip with sound.

  • Generates video from text alone, no image or footage needed.
  • Native audio comes with the clip.
  • Up to 30 seconds, 1080p, in the aspect ratio you choose.
Try Seedance 2.5 T2V

Image to Video

Turn a single image into motion.

  • Animates any still image into video with sound.
  • Set an end frame to control where the motion lands.
  • Keeps your image aspect ratio, up to 30 seconds.
Try Seedance 2.5 I2V

Reference to Video

Guide the video with your own references.

  • Use up to 50 references across images, videos, and audio.
  • Keeps characters, style, and sound consistent across shots.
  • Edit and extend existing clips on the same model.
Try Seedance 2.5 R2V
Read ByteDance's Seedance 2.5 prompting guide

Specs

Duration
30s4 to 30 seconds, single pass. Default 10s.
Resolution
1080p1080p by default, 480p available. Standard or high bitrate.
Aspect ratios
621:9, 16:9, 4:3, 1:1, 3:4, 9:16. I2V derives from your image.
Audio
NativeAudio-video generation, on by default.
Prompt characters
15,000Room for full shot direction, not one line.
R2V references
50Up to 30 images, 10 videos and 10 audio clips.

Features

  • 01

    Sound built in

    Native audio-video generation means dialogue, ambience, and effects come with the clip. No separate sound pass.

  • 02

    Long-form takes

    Up to 30 seconds in a single generation for real storytelling, not just 5-second tests.

  • 03

    Cinematic and photorealistic

    Tuned for film-grade motion and lighting.

  • 04

    Follows direction

    Stronger multilingual instruction following, including directional camera language.

  • 05

    Consistency across shots

    Reference-driven generation keeps characters and style locked, then edit and extend without starting over.

  • 06

    Anonymous by default

    Generated on Venice with no training on your inputs.

Lumara Film Festival

Now put it to work

Lumara

Everything above was built for exactly this. Lumara is an independent film festival for the artists shaping the next era of cinema, presented by MoonPay and Venice. Seven awards, $100,000 in cash prizes and matching Venice credits. All visual footage is AI-generated, on any platform you choose.

Bring a story with a clear point of view and the nerve to take risks.

Submit your film

Submissions close September 15, 2026.

Start creating with Seedance 2.5

Anonymous on Venice.

Start creating