Capability Map
Model Capabilities
bach-1.0-preview
bach-1.0-preview is the default video generation model, supporting three generation modes with configurable output parameters.
Supported Generation Modes
| Generation Mode | Duration | Resolution | Frame Rate |
|---|---|---|---|
| Text-to-Video | 1–6 s | 720p, 1080p | 24fps, 30fps |
| Image-to-Video | 1–6 s | 720p, 1080p | 24fps, 30fps |
| Multi-Image-to-Video | 1–6 s | 720p, 1080p | 24fps, 30fps |
Output Specifications
| Parameter | Value |
|---|---|
| Video Codec | H.264 |
| Container Format | MP4 |
| Frame Rate | 24 fps, 30 fps |
| Supported Aspect Ratios | 16:9, 9:16, 1:1 |
Text-to-Video
Generate videos directly from natural language prompts. This mode produces a 6-second video and is ideal for conceptual visualizations, creative storytelling, and rapid prototyping.
- Endpoint:
POST /videos/text2video - Duration: Configurable from 1 to 6 seconds
Image-to-Video
Animate a single static image guided by an optional text prompt. The model analyzes the source image and generates smooth, natural motion while preserving visual consistency with the input.
- Endpoint:
POST /videos/image2video - Duration: Configurable from 1 to 6 seconds
Multi-Image-to-Video
Compose a video from multiple reference images with optional subject reference synthesis. This mode supports complex scene compositions and maintains consistent character or object appearance across frames.
- Endpoint:
POST /videos/multi2video - Duration: Configurable from 1 to 6 seconds
- Maximum Images: 9 per request
On this page