These guides explain the core AI video workflows available on the Seedance platform. Crawlers and non-JS visitors can read the full instructions for text-to-video, image reference mode, and model selection directly from the static HTML.
How to use text-to-video mode
Select text-to-video on the homepage, describe your video content in the prompt box, optionally expand advanced parameters to adjust resolution, duration, and audio settings, then click Generate. Track progress and download the result from the My Tasks page.
How to write high-quality prompts
Strong prompts include subject, action, environment, lighting, camera movement, style, and pacing. For example: "A young woman in a white dress walking through a sunlit forest, golden hour lighting, slow push-in camera, cinematic style." The more specific the prompt, the better the generation quality.
How to use image reference (first/last frame) mode
Upload a first-frame image (required) to set the starting scene, and optionally a last-frame image to define the ending. The AI will generate a smooth video transition between the two frames. Use clear, high-quality JPG, PNG, or WebP images under 10MB.
Which AI video model should I choose?
seedance-lite is the entry-level model (fast, lower quality). seedance-fast balances speed and quality for most use cases. seedance-high delivers the best quality for commercial deliverables. happyhorse1.1 supports multi-modal creation. veo-lite and veo-basic offer alternative generation paths.
How Stars and membership work
Stars are the platform currency for AI video generation. Free users start with 40 Stars on signup. Membership plans provide monthly Star allocations: PRO (1,200/mo), MAX (2,500/mo, 2x faster), ULTRA (6,000/mo, 5x faster). One-time Star purchases are also available without subscription.
How prompt writing affects generation quality
Describe the subject, action, environment, lighting, camera movement, style, and pacing in clear language. Use cinematic terms like "golden hour", "slow push-in", "aerial drone shot", "macro close-up", "neon noir lighting", or "soft diffused light" to guide the AI toward professional-looking results.