kling-v1 is Kuaishou Kling's first-generation video generation model, suitable for turning text concepts or static images into short shots and further extending generated videos. Its practical features include support for five-second first and last frames and camera movement control, making it easy to arrange the start and end of scenes and camera direction. For creations that do not require native audio, 4K, or multi-material editing, this model can be used to build a streamlined video workflow.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
The above are the calling specifications for kling-v1 on this platform; photo lip-sync is a combined creation feature and does not mean the model natively generates audio.
When no existing visual is available, use text to describe the subject, action, and environment for text-to-video generation; when product or scene images are available, use an image as the first frame for image-to-video generation. The two methods are suited to concept exploration and animating an established visual, respectively, and output video links for continued editing.
Five-second image-to-video generation can specify both the first and last frames, giving the shot a clear starting point and endpoint; five-second tasks can also enable camera movement control. In practice, first determine the composition, then describe the intermediate action, treating shot direction as a separate control rather than relying only on a general prompt.
After obtaining a satisfactory clip, save the returned video_id, use extend, and add the next prompt to continue generation. Extensions are based on an existing generated video and are not equivalent to arbitrary video editing. Applications can use task_id to track tasks and receive video results through asynchronous queries or callbacks.
Provide a product image as the first frame, describe the background atmosphere and desired actions, and create a short showcase shot. When controlling the beginning and ending is needed, use the five-second first-and-last-frame option; when camera movement is needed, also choose a five-second task. The delivered video can be used as advertising edit material, with captions and music added in later production.
Write the subjects, scenes, and actions in a storyboard as prompts, then generate five- or ten-second clips to observe the concept in motion. Landscape, portrait, and square aspect ratios can meet different layout requirements. Once a satisfactory shot is available, try extending it and organize the material clip by clip instead of requesting a complete long video at once.
Provide a clear front-facing photo of one person and pre-recorded audio, then use the talking photo feature to create a short talking-head video. When calling it, explicitly select kling-v1; prompts are used to describe actions or expressions during the animation stage. The result includes the final lip-sync video, and an intermediate photo animation video can also be obtained.
If the task focuses on animating still images and requires five-second first-and-last frames or camera movement control, kling-v1 has a clear feature combination. Although kling-v1-6 is a later version, it does not support camera movement and cannot be directly substituted simply because its version is newer. Selection should center on camera control requirements rather than treating the version number as the sole basis for judgment.
If you need to generate audio while generating video, consider the pro mode of kling-v2-6 or kling-v3; if you need 4K and more flexible durations, consider kling-v3. For multi-image reference or editing existing videos, choose kling-o1 or kling-v3-omni. v1 is better suited to basic creation using text, first frames, and short camera control.
For text shots, select action=text2video; for image animation, select action=image2video and provide start_image_url. Clearly describe the subject's action and the scene, and start with one coherent shot.
Specify model=kling-v1 to /kling/videos, and choose std/pro and 5 or 10 seconds. Five-second image-to-video can use end_image_url, and five-second tasks can use camera_control according to the documentation.
Use async or callback_url to track tasks, and save the task_id and video ID; check the subject, action, and first-to-last-frame continuity. This model does not generate audio, so add sound in post-production; use the corresponding video ID when extending.
Start from a first frame of a person sitting on a chair and end with a last frame of the person standing up, maintaining the indoor composition, with slow, natural movement and a fixed camera.
Use 5-second image2video and check the starting and ending frames and body movement; do not reuse five-second-specific last-frame or camera movement controls for 10-second tasks.
Choose five seconds when first/last frame or camera movement control is needed; this is a clear functional boundary of this model. Ten seconds is suitable for generation tasks that do not depend on these two controls. The duration should be determined by the shot content; if you need to continue an existing clip, use video extension rather than assuming arbitrary lengths are supported.
No. Image-to-video requires start_image_url; the last frame, end_image_url, is only an optional ending reference and must be used together with the first frame. When using first and last frames with kling-v1, you must also choose a five-second task; when only using the first frame, describe the desired action around that image.
For initial generation, cfg_scale can adjust prompt relevance, ranging from 0 to 1; higher values adhere more closely to the prompt. negative_prompt can specify content you do not want to appear, up to 200 characters. Neither applies to extend; when continuing, focus on describing the action in the next segment.
Use /kling/talking-photo, submit image_url and audio_url, and explicitly set model to kling-v1. Supported audio formats are mp3, wav, m4a, and aac, with a maximum size of 5MB; videos can be five or ten seconds. It synchronizes lip movements using existing audio and does not create dialogue on its own.
Video tasks return information such as video_url, video_id, task_id, and status: the video link is used to retrieve the finished video, the video ID is used for subsequent extension, and the task ID is used to track processing progress. Setting async=true lets you obtain the task ID first and then query the result; you can also configure a callback address to receive completion information.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗Pricing
$0.13–$0.93· per second
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs