Kling-O3 - Poe

New chat

Kling-O3

@empiriolabsai

78 followers

API

Try Kling-O3

Kling O3 is a versatile AI video generation model capable of creating high-quality videos in Standard or Pro modes, with resolution varying by aspect ratio. It supports multiple workflows including Text-to-Video, Image-to-Video, Reference-to-Video, and Video Editing, with advanced features like native sound generation and multi-scene transitions. Multi-Scene Mode - Separate scenes with `|` in your prompt — automatically detected, no toggle needed (max 6 scenes) - Not compatible with video input workflows (video edit, reference with video) - Optional per-scene durations: `5s: Scene one | 3s: Scene two | 7s: Scene three` - Explicit per-scene durations override the total duration set - If no durations specified, the slider duration is distributed evenly across scenes - Per-scene duration may sometimes not be exact as the model treats them as guidance, not exact frame cuts Input Limits & Requirements - Prompt length: 3 characters minimum, 2,500 characters maximum - Image inputs: JPEG, PNG, or WEBP only. Max 10MB per file. Images below 300x300 are automatically upscaled; images outside the 1:2.5-2.5:1 aspect ratio range are automatically center-cropped to fit - Video inputs: MP4, MOV, or WEBM only. Videos longer than 10 seconds are automatically trimmed; videos outside the 720-2160px resolution range are automatically resized. Minimum 3 seconds (cannot be extended) - Video workflow duration: When a video input is used (Video Edit, Reference with video), output duration is automatically capped at 10 seconds - Image-to-Video: Max 2 images (first frame + optional last frame) - Reference-to-Video: Max 7 reference images (max 4 if a reference video is also included), max 1 reference video - Video Edit: Exactly 1 video, no images This bot supports optional parameters for additional customization.

OFFICIAL