Generate shots from text descriptions
Use action=text2video, explicitly specify model=kling-v3-turbo, and describe the subject, action, composition, and sound requirements in the prompt.
Overview
API
Pricing
Kling 3.0 Turbo · Video Generation
Generate short videos with native audio from text or a first-frame image
Use kling-v3-turbo to generate 3–15 second video clips. Choose std 720p or pro 1080p output, and structure prompts around the subject, action, scene, and sound.
Capabilities
Start from a first-frame image
Use action=image2video and start_image_url to make the input image the starting point of the shot, then describe the subject's movement and scene changes.
Native audio generated with the visuals
This model includes native audio and does not provide a switch to disable it. Omit generate_audio or set it to true; after generation, you should still verify that the audio meets delivery requirements.
Core Specifications
- Invocation ID
- kling-v3-turbo
- Endpoint
- POST /kling/videos
- Generation Methods
- Text-to-video, first-frame image-to-video
- Duration
- 3–15 whole seconds
- Quality
- std: 720p; pro: 1080p
- Audio
- Native audio, no switch to disable it
Use Cases
Individual shots in short films
Fit one clear action into 3–15 seconds, validate the visuals and sound first, then combine them into a longer work.
Animate images
Use images of people, products, or scenes as the first frame, and describe changes around the subjects already present in the image.
Sample clips with sound
Use real business prompts to compare quality, motion, and audio before deciding whether to use the generated result. The API does not guarantee that a single generation is ready for direct delivery.
How to Choose a Model
Choose Turbo for this input and audio combination
When the task only requires text or a single first-frame image, and native audio is acceptable, you can start with kling-v3-turbo. std and pro correspond to different output resolutions; see Pricing for actual billing.
Check specialized capabilities of other models first
If you need an end frame, camera movement objects, an independent negative prompt, or cfg_scale, you should not apply these parameters directly to Turbo. For multi-image references and existing video editing, choose the appropriate specialized model and operation, and verify against that endpoint's documentation.
Get Started
Prepare Your Goals and Materials
Determine the shot content and integer duration; for image-to-video, prepare an accessible start_image_url.
Explicitly Select the Model and Quality
Set model=kling-v3-turbo, choose std or pro as needed, and use the corresponding action.
Wait for Task Completion
You can set async=true to obtain a task_id and then query the result, or configure callback_url to receive completion notifications. Successful submission does not mean video generation is complete.
Usage Limits
- End frames, camera_control, standalone negative_prompt, and cfg_scale are not supported. Content to avoid should be included in the prompt.
- std outputs 720p and pro outputs 1080p; do not promise 4K configurations from other models for this model.
- Native audio does not provide a toggle to disable it; do not send generate_audio=false. Muted videos require post-processing.
- Duration must be an integer from 3–15 seconds. Output quality and generation time should be evaluated using actual samples; do not infer speed guarantees from the model name.
Frequently Asked Questions
How do I get started with text-to-video?
Submit action=text2video, model=kling-v3-turbo, and prompt to POST /kling/videos, and set a valid quality and integer duration.
What is required for first-frame image-to-video?
Use action=image2video and provide start_image_url. This model supports first-frame guidance but does not support end_image_url end frames.
What durations and resolutions can I choose?
Duration must be an integer from 3–15 seconds; std is 720p and pro is 1080p.
Can I disable generated audio?
This model includes native audio and does not provide a toggle to disable it. Omit generate_audio or set it to true; for muted output, post-process the completed video.
Does it support camera control or standalone negative prompts?
camera_control, negative_prompt, and cfg_scale are not supported. Describe shot requirements and content to avoid in the prompt.
How do I retrieve asynchronous task results?
After setting async=true, first obtain the task_id, then query the status and final video result through the task API, or use callback_url to receive completion notifications.
Platform API contract · Updated: 2026-10-01.
Evaluate Kling 3.0 Turbo with Real Tasks
Verify parameters and pricing before calling, and check the visuals and audio after generation.
Pricing
$0.11–$0.13· per second
Official $0.11–$0.14−4%vs official
Top-up
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.
Quickstart
Try the API
API Key
Please obtain an API Key first
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs