Video Generation - MyTokenGate

1. Introduction

Video generation models create a short video from a text description (text-to-video), drive a video from a reference image plus text (image-to-video), hold a subject and style steady across several reference images (reference-to-video), or reinterpret and restyle a source video (video-to-video and video editing). Common uses include short-form clips, concept demos, animated covers, and product showcases.

Generation is asynchronous: create a job to get a task ID, poll its status, and receive the video URL once it completes. The download URL returned after generation is a MyTokenGate address. It is retained for about 7 days, so download and keep your video soon after generation; after it expires you will need to generate again. See the Video Generation API for full details.

2. Supported Models

Four model families are available, covering five modes: text-to-video, image-to-video, reference-to-video, video-to-video, and video editing.

Within the Seedance family: iterate on your prompt and composition with seedance-2.0-mini, then switch to seedance-2.0 for the high-quality final render.

For a per-model table of modes and resolutions, plus request examples for each family, see the Video Generation API.

3. Prompt Tips

Describe actions and scenes in chronological order with specific detail: actions, appearance, camera angle, and environment, written cohesively in a single paragraph that starts directly with the action. Think of yourself as a cinematographer describing a shot script, and keep the prompt within 200 words.

A recommended structure:

• Start with one sentence for the main action. Example: A woman with light skin, wearing a blue jacket and a black hat with a veil.
• Add specific actions and gestures. Example: She first looks down and to her right, then raises her head back up as she speaks.
• Describe the character or object precisely. Example: She has brown hair styled in an updo, light brown eyebrows, and a white collared shirt under her blue jacket.
• Set the background and environment. Example: The background is out of focus, but shows trees and people in period clothing.
• Specify camera angle and movement. Example: The camera remains stationary on her face as she speaks.
• Describe lighting and color. Example: The scene is captured in real-life footage, with natural lighting and true-to-life colors.
• Note any change or event. Example: A gust of wind blows through the trees, causing the woman’s veil to flutter slightly.

4. Image-to-Video

Image-to-video adds a reference image on top of the text prompt. The image can serve as the first frame, or as a reference the model extends into later frames.

For stable results, use an image with a 16:9, 9:16, or 1:1 aspect ratio; the output resolution matches the image aspect ratio. The image must be reachable via a public HTTPS URL.

To keep the same subject or style across the whole clip, use reference-to-video: send several images at once, the first as the first frame and the rest as references. happyhorse-1.1-r2v and kling-o3-standard-reference support this. See the Video Generation API for the exact request format.

5. Video-to-Video

Video-to-video adds a reference video on top of the text prompt; the model reinterprets or restyles it while preserving the original motion and timing. It suits stylized adaptations, refreshing old footage, and similar work.

The reference video must be reachable via a public HTTPS URL, ideally close in length to the target output. Because it includes a video input, it usually consumes more tokens than text-to-video at the same resolution. Supported on seedance-2.0, kling-o3-standard-v2v, and kling-o3-pro-v2v.

Video editing is closely related: it also takes a source video as input and rewrites the picture according to a text instruction, via happyhorse-1.0-video-edit, kling-o3-standard-video-edit, and kling-o3-pro-video-edit. The request shape is identical — a source video block inside content. See the Video Generation API for the exact request format.

6. Billing

Video generation is billed by output tokens and settled at the actual usage when the job completes. If generation fails, the pre-authorized amount is released automatically with no charge. Pricing varies by model and resolution (for example, 1080p); see Billing for details.

7. Try It and Call It

• Online: sign in to the console and open the Video tab in Playground to generate from a prompt.
• API: see the Video Generation API for a full create-and-poll example.