Kling AI Face Model

Train the AI to Create Your Model

For those who are new to this feature, Kling AI’s Face Model feature basically lets you make a custom model of yourself that you can put into AI videos. It creates a virtual, lifelike version of you, using video data that you upload to the Kling AI servers, and you can then use that model whenever you want to make new AI videos.

The first part of this process involves training the AI so that it can build your custom model. To do this, you need to record a series of videos – between 10 and 30 in total – to upload to the Kling AI system. The platform will guide you and show you the sorts of movements and expressions to make, and the better quality your videos, the more detailed your model will be.

Video Upload Guidelines for Face Model Training

• Single Subject: Each video must feature only one person and should be recorded in one session to maintain consistent appearance (e.g., age, weight, makeup).
• Duration and Quality: Videos should be 5-15 seconds long and shot in 1080p resolution.
• Framing: The subject's face should occupy 1/18 to 1/4 of the frame, with clear visibility of facial features. Keep the camera focused on the face throughout.
• Action and Expression: Each video should showcase a single action or expression performed naturally at normal speed. Repeating the action at intervals is allowed.
• Filming Setup: Avoid handheld selfies. Use a tripod or have someone else film for stability. Consider varying the background slightly (e.g., changing camera angles) while keeping the lighting consistent—natural and balanced.
• Camera Stability: Use a fixed or stabilized camera to prevent shaking.
• Facial Consistency: Ensure the left and right sides of the face are consistent across clips. For example, if a mole is on the left side, it should appear there in every video. Adjust mirrored footage if filming selfies.
• No HDR Videos: Please do not upload HDR (High Dynamic Range) videos for training.
• Makeup: Makeup is optional but should be consistent across all clips.

Insert Yourself into AI Videos

Once you have made your own custom model through Kling AI’s Face Model feature, you can start using it in your next AI video creations. When you go to the text to video page, for example, you’ll see a new box appear that lets you add your model to the video, and you can type in a text prompt of whatever you want to see your model doing.

Customize Your Content

When making new AI videos with your custom face model, you’re free to use all of the usual customization features that Kling AI usually provides. That means you can choose from different durations and aspect ratios when generating your video, for example, and you can set the camera controls as desired and even add a negative prompt for an extra layer of control.

Youtube Videos

Other Reviews on X