Browse ModelsKwaivgiKwaivgi Kling Video O3 Pro Video Edit

Kwaivgi Kling Video O3 Pro Video Edit

Kwaivgi Kling Video O3 Pro Video Edit

Playground

Try it on WavespeedAI!

Kling Omni Video O3 Video-Edit enables conversational video editing through natural language commands. Remove objects, change backgrounds, modify styles, adjust weather/lighting, and transform scenes with simple text instructions like ‘remove pedestrians’ or ‘change daytime to dusk’. Ready-to-use REST API, best performance, no coldstarts, affordable pricing.

Features

Kling Video O3 Pro Video Edit

Kling Video O3 Pro Video Edit is Kuaishou’s most advanced video editing model, enabling natural-language-driven edits on existing video footage. Upload a video, describe the change you want — swap objects, alter scenes, shift styles — and get high-quality edited results with preserved motion and structure. Supports up to 4 reference images for precise visual guidance and optional original audio retention.


Why Choose This?

  • Prompt-driven editing Describe your edits in plain language — no timeline, no masks, no manual keyframing required.

  • Reference image support Attach up to 4 reference images to guide the target element, scene, or style in the output.

  • Audio preservation Keep the original soundtrack intact with the keep_original_sound option.

  • Element list control Lock in specific visual elements — characters, objects, or styles — to maintain consistency across the entire clip.

  • Scene-level understanding The model recognizes objects, backgrounds, and context within the video to apply accurate, context-aware edits.

  • Motion-consistent output Edits blend naturally across frames with strong temporal coherence — minimal flicker or ghosting.


Parameters

ParameterRequiredDescription
promptYesText description of the desired edit.
videoYesInput video to edit (URL or upload).
imagesNoUp to 4 reference images for element, scene, or style guidance.
keep_original_soundNoWhether to keep the original sound from the video. Default: enabled.
shot_typeNoEditing mode: intelligent (default, auto-determines edit scope) or customize.
element_listNoList of visual elements to maintain consistency throughout the clip.

How to Use

  1. Upload your video — drag-and-drop, file upload, or paste a public URL.
  2. Write your prompt — describe exactly what should change (e.g., “Change the man’s clothes in Figure 1.”).
  3. Add reference images (optional) — attach up to 4 images to steer the look of elements or styles.
  4. Set audio preference — toggle keep_original_sound to preserve or remove original audio.
  5. Select shot_type (optional) — use intelligent for automatic edit scope, or customize for manual control.
  6. Add element list items (optional) — see Notes below for how to use elements effectively.
  7. Run — submit and download the edited video.

Pricing

DurationCost
3s (minimum)$0.504
5s$0.840
10s (maximum)$1.680

Billing Rules

  • Base rate: $0.84 per 5 seconds ($0.168 per second)
  • Minimum billed duration: 3 seconds
  • Maximum billed duration: 10 seconds

Best Use Cases

  • Social Media Campaigns — Quickly swap products, backgrounds, or props in short-form videos.
  • Brand & Marketing — Replace or update branded elements across video assets without reshooting.
  • E-commerce — Edit product videos to showcase different variants, colors, or settings from a single source clip.
  • Creative Exploration — Experiment with style changes, scene swaps, and visual concepts on existing footage.
  • Storytelling & Film — Adjust scene details, atmosphere, or objects to refine narrative visuals in post-production.

Pro Tips

  • Use clear, specific prompts describing exactly what should change for best results.
  • Reference images work best when they clearly represent the target element or style.
  • Keep keep_original_sound enabled when audio continuity matters for your project.
  • Test edits on shorter clips first, then apply to longer footage once satisfied.
  • Ensure video URLs are publicly accessible — a preview thumbnail in the interface confirms the link works.

Notes

  • Both prompt and video are required fields.
  • Billed duration is clamped between 3 and 10 seconds regardless of actual video length.
  • Using element_list: First use Kling Elements to generate your element and note its name and ID. Then simply write the element name naturally in your prompt, and enter the corresponding element ID in the element_list field. No special characters or syntax required.
  • If using a URL, make sure it is publicly accessible.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result


# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/kwaivgi/kling-video-o3-pro/video-edit" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{
    "keep_original_sound": true,
    "shot_type": "customize"
}'

# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"

Parameters

Task Submission Parameters

Request Parameters

ParameterTypeRequiredDefaultRangeDescription
promptstringYes-The positive prompt for the generation.
videostringYes-The video URL. Video duration can not be longer than 10s.
imagesarrayNo[]-Including reference images of the element, scene, style, etc. Max 4
keep_original_soundbooleanNotrue-Whether to keep the original sound from the video.
shot_typestringNocustomizecustomize, intelligentShot type for the generation.
element_listarrayNo--Element reference list.

Response Parameters

ParameterTypeDescription
codeintegerHTTP status code (e.g., 200 for success)
messagestringStatus message (e.g., “success”)
data.idstringUnique identifier for the prediction, Task Id
data.modelstringModel ID used for the prediction
data.outputsarrayArray of URLs to the generated content (empty when status is not completed)
data.urlsobjectObject containing related API endpoints
data.urls.getstringURL to retrieve the prediction result
data.has_nsfw_contentsarrayArray of boolean values indicating NSFW detection for each output
data.statusstringStatus of the task: created, processing, completed, or failed
data.created_atstringISO timestamp of when the request was created (e.g., “2023-04-01T12:34:56.789Z”)
data.errorstringError message (empty if no error occurred)
data.timingsobjectObject containing timing details
data.timings.inferenceintegerInference time in milliseconds

Result Request Parameters

ParameterTypeRequiredDefaultDescription
idstringYes-Task ID

Result Response Parameters

ParameterTypeDescription
codeintegerHTTP status code (e.g., 200 for success)
messagestringStatus message (e.g., “success”)
dataobjectThe prediction data object containing all details
data.idstringUnique identifier for the prediction, the ID of the prediction to get
data.modelstringModel ID used for the prediction
data.outputsstringArray of URLs to the generated content (empty when status is not completed).
data.urlsobjectObject containing related API endpoints
data.urls.getstringURL to retrieve the prediction result
data.statusstringStatus of the task: created, processing, completed, or failed
data.created_atstringISO timestamp of when the request was created (e.g., “2023-04-01T12:34:56.789Z”)
data.errorstringError message (empty if no error occurred)
data.timingsobjectObject containing timing details
data.timings.inferenceintegerInference time in milliseconds
© 2025 WaveSpeedAI. All rights reserved.