Note: Kling v3 is available in both Standard and Pro tiers with dedicated text-to-video and image-to-video endpoints. Learn more in the official docs: Kling v3 Standard (T2V), Kling v3 Standard (I2V), Kling v3 Pro (T2V), and Kling v3 Pro (I2V).
We've added Kling v3 Standard and Kling v3 Pro as new video model options. Both models support text-to-video and image-to-video, and the platform now automatically selects the correct Kling v3 endpoint based on whether a start image is provided.
What's New
- Two New Models: Added
kling-v3-standardandkling-v3-proto the video model catalog. - Automatic I2V/T2V Routing: The integration now automatically routes to image-to-video when a start image is present, and to text-to-video when it is not.
- Duration Range: Supports 3 to 15 seconds.
- Aspect Ratios: Supports 16:9, 9:16, and 1:1.
- Optional Native Audio: Supports Kling v3 native audio generation with optional voice control.
About Kling v3 Integration
Both Kling v3 tiers are wired end-to-end across provider schemas, generation service routing, model metadata/pricing, and Mastra tool mapping.
- Single model selection in product UI: choose either Standard or Pro.
- Mode handled automatically: text-only prompt uses T2V; prompt + first frame uses I2V.
- Consistent metadata handling: duration, aspect ratio, image mode, and audio flags are captured in output metadata.
Pricing
Pricing is now available for both tiers in the model metadata and generation pipeline:
- Kling v3 Standard: audio-off and audio-on pricing, with voice-control pricing applied when voice IDs are provided.
- Kling v3 Pro: audio-off and audio-on pricing, with voice-control pricing applied when voice IDs are provided.
How It Works
Select Kling v3 Standard or Kling v3 Pro in the video model selector. Provide:
- a prompt only for text-to-video, or
- a prompt plus first-frame image URL for image-to-video.
The platform will route to the correct endpoint automatically and return an MP4 output with normalized duration/aspect handling.