Cinematic Text-to-Video up to 30 Seconds with Seedance 2.5

Seedance 2.5 is ByteDance's newest AI video generation model. It creates cinematic videos up to 30 seconds long from text prompts, with synchronized audio — dialogue, sound effects, and background music — generated by default. On Scenetra it comes in three modes: Text to Video, Image to Video, and Reference to Video, all usable as nodes in the visual workflow editor.

Yes — Seedance 2.5 is live on Scenetra right now, in all three modes. You don't need an API key, a waitlist, or a separate app: sign up, drop a Seedance 2.5 node onto your board, write a prompt, and generate. Scenetra is web-based and works anywhere.

Seedance 2.5 is a paid model billed per second of video, but you can try it without paying: new Scenetra accounts start with free welcome credits and a 7-day trial, which covers your first generations. After that you pay per generation from your credit balance — there is no subscription tier gating the model and nothing to cancel.

On Scenetra, Seedance 2.5 costs about $0.14 per second at 480p and $0.30 per second at 720p, audio included. A 5-second 720p clip costs about $1.51; a full 30-second clip about $9.03. You only pay for what you generate.

Any whole-second duration from 4 to 30 seconds, generated in a single pass. That makes it the longest-form model in the Seedance family — Seedance 2.0 tops out at 15 seconds.

Yes — audio is on by default. The model generates synchronized voice, sound effects, and background music together with the picture, so dialogue is lip-synced without any post-processing. You can disable audio generation with a toggle if you need silent footage.

They trade blows: Seedance 2.5's edge is 30-second single-pass generations with native synchronized audio and heavy multimodal referencing (up to 30 images, 10 videos, 10 audio tracks). Sora 2 and Kling 3.0 are stronger picks for some styles and offer higher resolutions. On Scenetra you don't have to choose — all of them are available side by side on the same board, so you can run the same prompt through each and compare.

You can feed existing footage in as reference: Reference to Video accepts up to 10 reference videos (30 seconds combined) for motion, style, and continuation-style generations. The model's dedicated frame-level video-editing mode is not yet exposed on Scenetra.

Three headline upgrades: durations up to 30 seconds (vs 15), audio-first generation, and a high-fidelity mov output option for editing pipelines. In Reference to Video mode it also accepts far more reference media — up to 30 images and 10 videos vs 9 and 3. Seedance 2.0 still offers a 1080p tier, while 2.5 currently caps at 720p.

480p and 720p, at aspect ratios 16:9, 4:3, 1:1, 3:4, 9:16, and 21:9 — or 'adaptive', which picks the ratio automatically from your prompt context. At 480p, 16:9 renders 854×480 and 9:16 renders 480×854.

Yes — use Seedance 2.5 Image to Video to animate a first-frame image (with an optional last frame to control the ending), or Reference to Video to guide generations with up to 30 reference images, 10 videos, and 10 audio tracks cited in the prompt as @Image1, @Video1, @Audio1. Both are available as separate nodes on your Scenetra board.