Kling AI upgrades motion control in VIDEO 2.6

Getting an AI video character to pull off a fast martial arts combo or a complex dance routine usually ends in melted fingers and warped faces.

Kling AI just rolled out an upgraded Motion Control feature inside Kling VIDEO 2.6, promising full control over every action and expression.

Bring your own moves

The update focuses heavily on fast, complex full-body motions like dancing and fighting, where generative models typically struggle the most.

According to Kling AI, the new system delivers precise hand gestures without blur, alongside natural facial expressions and accurate lip sync.

Creators can now upload a 3 to 30 second reference video to drive continuous, uninterrupted action sequences, while still using text prompts to fine-tune scene details.

Why it matters

AI video is quickly shifting from random generation to predictable, director-level control.

For animators, indie filmmakers, and marketers, getting a character to perform an exact physical movement used to require expensive motion capture or endless rerolls. Dropping in a phone-recorded 30-second clip as a direct motion guide makes real production pipelines far more viable.

We think targeting fast movements and hand clarity tackles AI video's two biggest immersion killers. It looks like a huge win for workflow consistency, though creators will want to test how well it holds up under tricky camera angles and heavy occlusions.