MKT SoftwareDocsOpen the app

Clone video

Screenshot a few moments of a video you like; the model reads how it was shot and builds a new video in the same style from your images.

Steps

  1. 1

    Open Clone video

    Go to Create video, group Topic & content, and pick Clone video.

  2. 2

    Add screenshots of the sample

    Choose 2 to 5 screenshots of different moments of the sample video, in time order.

  3. 3

    Analyze the video

    Click Analyze video. The model returns a summary: genre, scene count, aspect ratio, and whether there is dialogue.

  4. 4

    Add your images

    Choose 1 to 5 images of your people, product or setting. The model picks the best fit as the opening frame of each scene.

  5. 5

    Customize and confirm

    Adjust aspect ratio, language, scene count and what should differ from the sample, read the credit breakdown and click Confirm & create video. The finished video lands in the video library.

Screenshotting the sample

The tool does not take a video file. It reads the sample through screenshots, so the more representative they are, the closer the clone.

  • Capture the start, middle and end — right at scene changes, never the same scene twice.
  • Keep the original framing; do not crop.
  • Avoid shots covered by player controls, system captions or notifications.

The analysis

The model writes a scene-by-scene script from the screenshots but keeps it hidden; you only see the summary. Creating the video means you accept building from that summary.

  • If the summary is off, change the screenshots and click Analyze again.
  • Analysis spends no credit.
  • An analysis is kept for 24 hours. After that, analyze again before creating the video.

Customizing the clone

  • Aspect ratio — same as the sample, landscape 16:9 or portrait 9:16.
  • Dialogue language — Vietnamese or English.
  • Scene count — prefilled from the sample, adjustable from 1 to 8.
  • What should differ from the sample? — optional, up to 300 characters. For example: swap the product for an air purifier, a male presenter, an office setting.

Your people and product stay as they are

Each scene starts from one of your images, so faces and products stay true to them. Sharp images with one subject each work best.

Duration and cost

Every scene is exactly 10 seconds, so the video runs scene count × 10 seconds. The tool is billed per second on that total and runs on Server 1 only. Credits are held when you confirm, charged when the video finishes and returned to your wallet if the render fails. See the rate in how credits work.

Notes

  • Only clone videos whose style you are allowed to reuse. The tool copies the camera work and pacing, not the original footage.
  • The web form renders and saves to the library; it does not publish. Publish from the video library.
  • AI assistants can use this tool over MCP: analyze_video_clone analyzes, create_video_clone renders the video.