How to Use Seedance 1.5 for AI Video Generation

Seedance 1.5 is ByteDance’s text-to-video model that produces 1080p clips up to 12 seconds with synchronized audio. You access it through the Dreamina web app or the Volcano Engine API, pick a resolution and aspect ratio, write a prompt that describes motion and camera work, then render. First renders take roughly 60-90 seconds.

What is Seedance 1.5?

It is a diffusion-based video generator from ByteDance, the company behind TikTok and CapCut. Unlike the first Seedance, version 1.5 adds lip-sync and ambient sound in the same pass, so you do not have to run a separate audio tool. The model handles text-to-video and image-to-video, and it understands a 512-token prompt window.

Output lands at 720p or 1080p, 24fps, in 5s, 8s, or 12s lengths. Vertical 9:16 and widescreen 16:9 are both supported, which matters if you post to Shorts or Reels.

How do I generate a video with Seedance 1.5?

1. Open Dreamina and sign in with a ByteDance, TikTok, or Google account.
2. Switch the model selector to Seedance 1.5 (it defaults to the older version on some accounts).
3. Enter a prompt. Be concrete: “slow dolly-in on a ceramic cup, steam rising, morning window light, shallow depth of field.”
4. Choose resolution (1080p uses more credits), duration, and aspect ratio.
5. If you are doing image-to-video, upload a still under 10MB as the first frame.
6. Hit Generate. A 5s 720p clip typically finishes in 60-90 seconds.

Keep the camera instruction and the subject in separate sentences. Mixing “a woman walking” with “rapid zoom” in one clause often produces a static shot, in my experience.

Is Seedance 1.5 free?

No, not for sustained use. Dreamina gives new accounts a small pool of trial credits that covers two or three 5s clips. After that you buy credit packs; a 1080p 5s render costs more than a 720p one. The API on Volcano Engine is billed per second of generated video, which is cheaper for volume work but needs a Chinese phone number for signup. If you only need a handful of clips a month, the pay-as-you-go web path is simpler.

Seedance 1.5 vs Wan 2.2: which should you pick?

Wan 2.2 is Alibaba’s open-weight alternative and runs locally if you have a 24GB GPU. That is its big edge: no per-render fee and full control. Seedance 1.5 wins on polish, audio, and speed without hardware. Pick Seedance for fast commercial clips; pick Wan 2.2 if you want to fine-tune and avoid credits.

Where does Seedance 1.5 fall short?

Two things stand out. Hands and fast action still break — a person waving can smear across frames. And the 12-second ceiling is tight for anything narrative; you will need to stitch clips, and cuts between them are not always seamless in lighting. Text rendered inside the video is also unreliable, so avoid on-screen titles as a creative choice.

Quick Summary

• Seedance 1.5 is ByteDance’s text/image-to-video model with native audio, up to 12s at 1080p.
• Access it via Dreamina (web credits) or the Volcano Engine API (per-second billing).
• Free trial credits cover only a few clips; sustained use is paid.
• Best for short product and social clips; weak on hands, fast action, and embedded text.
• Choose Wan 2.2 instead if you want a locally-run open model.