FireLaunchFireLaunch
Sign in
Seedance 2.5 logo

Seedance 2.5

Generate 30-second AI videos from prompts, images, video cli

AI video generator that creates up to 30-second videos from text prompts and multimodal references (images, videos, audio). Supports targeted editing, video extension, and prompting in 10+ languages.

Seedance 2.5 is a video generation model built by ByteDance's Doubao team. It creates continuous videos up to 30 seconds long from a text prompt combined with up to 50 reference assets—30 images, 10 video clips, and 10 audio files. You write a prompt (supporting 10+ languages), optionally upload reference files to guide characters, scenes, motion, and sound, then generate.

The core workflow covers narrative scenes, ads, product visualization, and educational content. Unlike shorter-form generators, the 30-second limit lets you build complete story beats and emotional arcs in a single generation. You can also extend existing footage for temporal continuity, use audio alone to guide pacing and lip movements, and make targeted edits to specific visual elements without regenerating the full video.

Compared to Seedance 2.0, this version doubles the max duration (15s to 30s), increases reference capacity (9 to 30 images, 3 to 10 videos and audio files), adds standalone audio workflows, and improves targeted editing. New users get free trial credits to test one 5-second video.

🜂 Spark
0

0/250 to The Forge · 2d left

Spark over time

Screens

Seedance 2.5 screenshot 1

Conversation

Be the first to weigh in.

Sign in to join the conversation.

No comments yet. Start the conversation.

As featured on