Kuaishou’s Kling AI opened Kling 4.0 to all users on its website on 16 October, Pandaily reported, moving it beyond preview access. A single generated clip can run for up to 30 seconds. Users can set up to 10 keyframes — points in the sequence that steer changes in character, scene or plot — and combine up to 10 images, five video clips and seven subjects in one generation job. The product also supports dual-channel stereo and lip sync across Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French and Hindi, along with several Chinese dialects and English accent variants.
Making a short video with spoken dialogue could therefore involve setting visual changes and generating the sound with the picture, rather than treating the two as separate parts of the task.
An earlier AI Affairs report on the preview covered the 30-second clip length and multi-reference controls. Kling AI says the model’s capabilities are now connected to MCP and command-line workflows, which allow developers to use them through other tools. It lists 4K and 1080p 10-bit HDR output and an enterprise API as coming soon.
.jpg){kind=link}