ComfyUI-KLingAI-OmniVideo
Five API nodes for KLingAI's OmniVideo (O1) usage
Nodes (5)
First frame, last frame — and Kling Omni does the boring part in between
Up to five reference images, one animated result
Kling Omni video from inside ComfyUI
Tell Kling to change the scene, not re-render it
Kling Omni video extend
ComfyUI-KlingAI-OmniVideo 🚀
ComfyUI-KlingAI-OmniVideo is an unofficial ComfyUI node pack for Kling AI’s latest Omni-Video (kling-video-o1) model. It wraps API calls, OSS uploads, and video assembly into standard ComfyUI nodes, supporting text-to-video, multi-image storytelling, start/end-frame control, video editing, and continuation.
Node Guide
<table> <tr> <td><img width="720" height="297" alt="Node overview 1" src="https://github.com/user-attachments/assets/db28097f-b1a2-4974-a813-09497c76396b" /></td> <td><img width="958" height="637" alt="Node overview 2" src="https://github.com/user-attachments/assets/52b154d8-fdad-4eb7-9c07-124820a45db5" /></td> </tr> </table>✨ Highlights
- Full coverage of the five core Kling Omni modes.
- Automatic OSS relay: uploads images/videos and provides public URLs to the API.
- Native ComfyUI I/O: IMAGE/LATENT (via decode)/AUDIO friendly.
- Prompt helpers: use
<<<image_1>>>placeholders to precisely reference inputs.
🛠️ Installation
Manual (recommended):
- Go to your ComfyUI custom nodes directory:
cd /your_path/ComfyUI/custom_nodes/ - Clone this repo:
git clone https://github.com/starsFriday/ComfyUI-KLingAI-OmniVideo.git - Install dependencies:
cd ComfyUI-KLingAI-OmniVideo pip install -r requirements.txt - Restart ComfyUI.
⚙️ Configuration (required)
This plugin relies on the Kling AI API and Aliyun OSS (for temporary uploads). Create config.ini in this folder (or use the existing one):
[AUTH]
; Kling AI access/secret keys
; Get them at: https://klingai.kuaishou.com/api
kling_access_key = your_kling_ak
kling_secret_key = your_kling_sk
[OSS]
; Aliyun OSS config (temporary storage for URL access)
; Use a RAM sub-account with RW permission on the bucket
oss_access_key_id = your_aliyun_ak
oss_access_key_secret = your_aliyun_sk
oss_endpoint = https://oss-ap-southeast-1.aliyuncs.com
oss_bucket_name = your_bucket_name
; Optional custom domain (fallbacks to endpoint if empty)
oss_custom_domain = https://your.static.domain.com
The OSS bucket must be public-read or fronted by CDN so Kling servers can fetch your URLs.
⚠️ FAQ
- Input limits: video duration must be ≤10s, and each side of any input video should be within 720–2560 px; otherwise the API returns errors.
- Task Failed: verify keys in
config.iniand account balance. - Black output or download errors: ensure the OSS bucket is public-read or otherwise accessible.
- Audio: outputs include audio; to preserve original audio, combine with nodes like
VHS_LoadVideo.
License
MIT License. This package only wraps the API; model rights belong to Kuaishou Kling AI.