Generating Videos

Generating Videos on imgmov

imgmov lets you generate short AI videos from text prompts or by animating existing images. Multiple video models are available in one interface.

Available video models

Model	Best for	Credit cost
Agnes Video V2	Short motion clips	0 (free)
CogVideoX-Flash	Image-to-video motion	0 (free)
HappyHorse	Creative stylized video	3
Video-01-Live	Live-action-style video	4
Kling v3	Smooth cinematic motion	5
Hailuo 2.3	Dramatic, cinematic scenes	5
Veo 3.1	Highest-end video quality	8

Video duration

Clips can be generated at 4 seconds or 8 seconds, depending on the model. Shorter clips render faster and cost fewer credits.

Image-to-video workflow

The most effective video workflow starts with a still image:

1. Generate or select a seed image. Use an image model (see Generating Images) or pick one from the gallery.
2. Upload the seed image as a first-frame reference in the video generation panel.
3. Describe the motion in your prompt. Focus on camera movement and subject action.
   • Example: “Slow camera pan left, koi swimming in the pond, cherry blossoms swaying gently.”
4. Generate. The video model animates your image according to the prompt.

Reference modes

For video generation, reference frames can be used in two ways:

• Universal: General visual reference. The model uses the image as a style/composition guide.
• First-last-frame: The image is used as the starting and/or ending frame of the video. Only some providers support last-frame references.

Resolution and quality

Video resolution tiers work the same as image: 1K (SD), 2K (HD), and 4K (UHD, Pro only). Quality tiers are Standard, Medium, and High.

Tips for better video

• Start with a strong still image. The better your seed image, the better your video.
• Keep motion descriptions simple and specific. Too many motion instructions can produce chaotic results.
• Use free models (Agnes Video V2, CogVideoX-Flash) for quick experiments before spending credits on premium models.