ドキュメントKling 3.0
Model
kling-3.0
Mode-based Kling model with single-shot and multi-shot request shapes.
Model supplement for /unified-video/quote and /unified-video/generate権限: 実際の backend API key scope 設定に依存します
向いている用途
Teams that want one richer Kling model with mode, audio, and shot logic.
std/pro mode pricing
Single-shot and multi-shot
Element references
What you can send
- Required: model
- Single-shot: prompt
- Optional: images, aspectRatio, duration, generateAudio, mode, multiShots, multiPrompt, klingElements, projectId
Field rules
aspectRatio
Supported: 16:9, 9:16, 1:1.
duration
Integer 3..15. Sent as a string.
mode
std or pro.
multiShots
false by default. When true, multiPrompt becomes required and prompt is ignored.
multiPrompt
Max 5 shots. Each shot needs prompt (max 500 chars) and duration (1..12).
generateAudio
When multiShots is true, defaults to true on upstream side.
images
Single-shot allows up to 2 (first/last frame). Multi-shot allows at most 1 (first frame).
klingElements
Max 3 elements. Each element requires name, description, and 2-4 input images.
Pricing
Mode
Audio Off
Audio On
std
12
17
pro
15
23
Frontend notes
- Show single-shot first and reveal multi-shot only when needed.
- Default sound to true when multiShots is on.
- Keep element references under advanced settings.
- Use callback as the main completion flow.
すぐに試したいですか？
Settings で general API key を作成してから、ここに戻ってリクエスト例をコピーしてください。