Kling video models on AI Gateway
Source: Vercel Blog
Overview
Kling video models are now available in AI Gateway, including the newest Kling 3.0 models. These state‑of‑the‑art models can generate cinematic videos from text, images, or motion references, and are accessible through both AI Gateway and the AI SDK.
- Video generation is currently in beta and available for Pro, Enterprise, and paid AI Gateway users.
Core Capabilities
| Feature | Description |
|---|---|
| Text‑to‑Video | Generate a video from a simple text prompt. |
| Image‑to‑Video | Animate still images into video clips. |
| Multishot Narratives | Create multi‑scene videos from a single narrative prompt; the model intelligently cuts between shots to tell a complete story. |
| Audio Generation | Synchronized sound effects and ambient audio are generated alongside the video. |
| First & Last Frame Control | Provide explicit start and end frames for precise scene transitions (ideal for time‑lapse effects). |
| High‑Resolution Output | Supports up to 1080p generation in Pro mode. |
| Programmatic Access | Generate videos via AI SDK 6 (generateVideo). |
| No‑Code Playground | Experiment with video models directly in the AI Gateway Playground without writing code. |
Example Workflows
1. Text‑to‑Video
Generate a video of a cherry blossom tree using only a text prompt.
Prompt: "A cherry blossom tree swaying gently in the spring breeze."
Model: klingai/kling-v3.0-t2v2. Multishot Narrative
Create a narrative video with multiple scenes from a single prompt. The model decides optimal shot composition and produces synchronized audio.
{
"prompt": "Scene 1: A sunrise over a quiet lake. Scene 2: A fisherman casts his line. Scene 3: The sun sets behind the mountains.",
"shotType": "multishot",
"sound": "on",
"model": "klingai/kling-v3.0-t2v"
}3. First & Last Frame Control
Provide explicit start and end images to control how the video begins and ends (useful for time‑lapse or precise transitions).
{
"firstFrame": "url/to/start_image.png",
"lastFrame": "url/to/end_image.png",
"model": "klingai/kling-v3.0-i2v"
}Getting Started
- Simple: Text‑to‑Video with Audio – use
klingai/kling-v3.0-t2v. - Advanced: Multishot Video – also
klingai/kling-v3.0-t2vwithshotType: 'multishot'. - Advanced: First & Last Frame Control – use
klingai/kling-v3.0-i2v.
For detailed configuration options and sample scripts, see the Video Generation Documentation and Video Generation Quick Start guides in AI Gateway.
Available Models
| Model ID | Type | Description |
|---|---|---|
klingai/kling-v3.0-t2v | Text‑to‑Video | Latest generation, highest quality with multishot support |
klingai/kling-v3.0-i2v | Image‑to‑Video (First‑and‑Last‑Frame) | Animate images with v3 quality and multiple frames |
klingai/kling-v2.6-t2v | Text‑to‑Video | Includes audio generation support |
klingai/kling-v2.6-i2v | Image‑to‑Video (First‑and‑Last‑Frame) | Use images as reference |
klingai/kling-v2.5-turbo-t2v | Text‑to‑Video | Faster generation |
klingai/kling-v2.5-turbo-i2v | Image‑to‑Video (First‑and‑Last‑Frame) | Faster generation |