Kling: kling-4.0/text-to-video

Kling 4.0 Text-to-Video is Kuaishou's text-only entry point to the Kling 4.0 family. It generates a 3 to 30 second clip from a written prompt alone — no source image or reference footage required — at 720p or 1080p resolution and in four aspect ratios: 16:9, 9:16, 1:1, and 21:9. Native stereo audio is rendered in sync with the picture, including two-channel dialogue, sound effects, and ambience, with improved lip sync for characters speaking in multiple languages. The model accepts either a single text prompt or a multi-shot storyboard, letting you script several shots with individual durations into one continuous video. Prompts of up to 26,000 characters can cover subject, motion, camera direction, and sound in one block, backed by prompt understanding of up to 8,000 tokens. A single generation task can draw on up to 15 reference assets — up to 10 images, 5 videos totaling up to 30 seconds, 7 elements, and 3 video elements — plus up to 10 keyframe inputs for defining key visual moments. It is built for short-form and social video, advertisements and product demos, multi-shot narrative content, and previsualization. Large-scale motion stays stable through push-ins, pull-outs, pans, tracking shots, and orbit shots, and output remains coherent during high-speed action. A Flash mode offers faster generation at lower cost with videos up to 20 seconds at maximum 720p resolution, though it does not support start-and-last frames or multiple keyframes.