Unifically LogoUnificAlly
Model logo

Kling 3.0 Omni API

Image to Video

Kling 3.0 Omni image-to-video with sound support. 5-10s at 720p-1080p.

Click or drag & dropPNG, JPG, WEBP, GIF · Max 100MB
Click or drag & dropMP4, WEBM, MOV · Max 100MB
Keep reference video audio
Keep audio from the reference video
Output

Your generated video will appear here

Features

What Kling 3.0 Omni API offers

Omni video from a text prompt with references: up to 7 images, optional IMAGE or VIDEO elements JSON, optional start and end frames, optional reference video
Single-shot duration 3 to 15 seconds, or multi-shot with 2 to 6 scenes totaling 3 to 15 seconds
Optional batch of 1 to 4 outputs per single-shot request
Aspect ratio 16:9, 9:16, 1:1, or Auto based on frames and references
Reference video with Reference or Transform behavior and optional keep audio from that clip
Standard (720p) or Pro (1080p) output
REST API with JSON request and response bodies

Use cases

Built for

Primary

Directed generation - Prompt with @image_1 and @element_1 style references for consistent subjects

#2

Shot lists - Multi-shot for several beats in one 3 to 15 second timeline

#3

Style and motion transfer - Reference clip in Reference or Transform mode to steer look or retime motion

#4

Variants - Request multiple outputs in one call when exploring layout and motion

#5

Social and web - Auto or fixed aspect ratio for feeds, stories, and landing page loops

FAQ

About Kling 3.0 Omni API

It is the omni video model: one prompt can reference uploaded images, optional elements, start and end frames, and an optional reference video. Duration is 3 to 15 seconds in single-shot mode, or multi-shot with 2 to 6 scenes over the same total length.

Omni adds richer conditioning: more reference images, elements, optional reference video with Reference or Transform, keep audio from that video, Auto aspect ratio, and counting multiple outputs. Base Kling 3.0 focuses on prompt, optional frame images, multi-shot, sound, and aspect presets without the omni reference bundle.

Use @image_1 through @image_7 for uploaded images, @video_1 for the reference video when provided, and @element_1 style tags for named elements from the elements JSON, as described in the docs.

Reference treats the clip as a guide for style or motion. Transform is aimed at stronger edits or retiming of the provided video behavior. When you keep audio, the soundtrack from the reference clip can be preserved.