JoyAI-Video-Edit
Real-Time Open-Ended Video Editing with Autoregressive Diffusion
🐶 JoyAI-Video-Edit
JoyAI-Video-Edit is a real-time, instruction-guided video editing system for open-ended video streams. Given a live camera stream or uploaded video and a natural-language edit instruction, it edits frames causally as they arrive, without waiting for the full video, requiring a predefined video length, or revisiting future frames. In our deployment benchmark, the full end-to-end pipeline reaches 30 FPS at 720 × 1248, pushing video editing from offline batch processing toward interactive streaming generation.
The system combines an MLLM-based condition encoder, a causal video VAE, and a 16B-parameter multimodal diffusion transformer. It is trained and deployed as an autoregressive diffusion editor, then accelerated with aligned autoregressive distribution matching distillation, long-horizon optimization, bounded KV-state inference, and deployment-oriented scheduling to sustain high-throughput 720p editing while reducing train-inference mismatch and accumulated temporal drift.
🔥 News
2026.10.09 — 🎉 We open-source our example training scripts and release LongV2VBench for long-form video-to-video editing evaluation.
2026.08.24 — 🎉 Consumer GPU support landed: real-time streaming video editing on a single GeForce RTX 5090 (32 GB) at 840 × 480 @ 24 FPS. Deployment Guide
2026.08.15 — 🎉 Live demo released — real-time streaming video editing on a single RTX PRO 6000 (Blackwell) GPU: 840 × 480 @ 24 FPS or 720p @ 16 FPS. Try HuggingFace Demo
2026.08.14 — 🎉 Released an upgraded checkpoint with significantly stronger reference-image-guided video editing (RV2V), delivering better subject and identity preservation, more faithful reference conditioning, and improved temporal consistency across long streams. Grab the new DiT weights.
2026.08.05 — 🎉 We released the model checkpoints, deployment code, and technical report.
💎 Highlights
- Real-time open-ended editing. Edits live or uploaded videos as frames arrive, without requiring the full sequence upfront.
- Diverse instruction control. Supports subject edits, local edits, background changes, style transfer, motion changes, and reference-guided editing.
- Autoregressive diffusion design. Combines an MLLM condition encoder, causal video VAE, and MMDiT backbone for streaming video editing.
- High-throughput 720p deployment. Reaches 30 FPS end-to-end throughput at 720 × 1248 with bounded KV-state inference and stable per-chunk compute.
📦 Model Files
| File | Purpose |
|---|---|
dit/joyai_video_edit_dit_0811.pth |
Current DiT weights — use this one (stronger RV2V, released 2026.08.14) |
dit/joyai_video_edit_dit_0804.pth |
Initial release (kept for reproducibility; the server does not use it) |
vae/ |
Causal video VAE (config + safetensors) |
Place them under deploy/deps/checkpoints/JoyAI-Video-Edit/ — see the
Deployment Guide
for the full layout and the exact hf download command.
🎬 Demo
Try our online real-time video editing demo:
👉 https://huggingface.co/spaces/wxDai/joyai-video-edit
Project repository:
👉 https://github.com/jd-opensource/JoyAI-Video-Edit
Technical report:
👉 https://arxiv.org/abs/2608.03974
📚 Citation
If JoyAI-Video-Edit is useful for your research or project, please cite:
@article{xiao2026joyai,
title={JoyAI-Video-Edit: Real-Time Open-Ended Video Editing with Autoregressive Diffusion},
author={Xiao, Yicheng and Dai, Wenxun and Qin, Xinran and Song, Lin and Zhang, Maoquan and Xu, Hang and Chen, Yukang and Li, Yitong and Zhang, Guohui and Zhang, Yuan and Zhang, Xuying and Zhang, Tommy and Yuan, Jianlong and Li, Peihao and Lu, Shuai and Fu, Siming and Zhao, Chuyang and Han, Xin and Huang, Jie and Li, Wenbo and Ma, Guoqing and Huang, Wei and Qi, Xiaojuan and Huang, Haoyang and Duan, Nan},
journal={arXiv preprint arXiv:2608.03974},
year={2026}
}
📄 License
JoyAI-Video-Edit is released under the Apache License 2.0.
Please refer to the project repository for the complete license:
https://github.com/jd-opensource/JoyAI-Video-Edit/blob/main/LICENSE