LongCat Video – Generate Minutes‑Long High‑Quality AI Videos

LongCat Video empowers creators with a unified LongCat AI pipeline. With LongCat‑Video you can run Text‑to‑Video, Image‑to‑Video, and Video‑Continuation locally to produce long, coherent, high‑quality content. Start building with Long Cat Video today.

See LongCat Video in Action

Watch how LongCat‑Video generates high‑quality long videos with LongCat AI locally.

LongCat Video Examples

Explore sample outputs generated by LongCat Video and LongCat AI. Each video showcases long, coherent motion and high visual quality from the LongCat‑Video model.

LongCat‑Video Technical Report

Preview the LongCat‑Video technical report below. Download the PDF or open it in a new tab to explore details on the LongCat Video architecture, LongCat AI training (multi‑reward GRPO), and efficient long‑video generation.

Your browser does not support inline PDF preview. Please open the LongCat‑Video report or use the download button above.

Quick Start

Set up LongCat‑Video locally and start generating long, high‑quality videos with LongCat AI.

Installation

Clone the repo:

git clone https://github.com/meituan-longcat/LongCat-Video
cd LongCat-Video

Create conda environment:

conda create -n longcat-video python=3.10
conda activate longcat-video

Install torch (configure according to your CUDA version):

pip install torch==2.6.0+cu124 torchvision==0.21.0+cu124 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu124

Install flash-attn-2:

pip install ninja
pip install psutil
pip install packaging
pip install flash_attn==2.7.4.post1

Install other requirements:

pip install -r requirements.txt

LongCat Video on X

How LongCat Video Works

Create with LongCat Video in three steps: pick a task, tune settings, and generate minutes‑long videos with LongCat AI.

1

Pick a Task Choose Text‑to‑Video, Image‑to‑Video, or Video‑Continuation in the LongCat‑Video playground.

2

Tune Parameters Set duration, resolution, and motion controls for your LongCat Video generation.

3

Generate & Share Render with LongCat AI, then download and share your Long Cat Video output.

Try Create AI Video Now

Core Features of LongCat Video

Discover how LongCat Video and LongCat AI make high‑quality, long video generation fast and reliable. LongCat‑Video unifies tasks while keeping text and image alignment strong.

LongCat Video FAQ

LongCat Video is a foundational, open‑source LongCat AI model for Text‑to‑Video, Image‑to‑Video, and Video‑Continuation. LongCat‑Video specializes in generating minutes‑long, coherent videos at 720p 30fps.
LongCat Video uses a unified dense architecture with coarse‑to‑fine generation along temporal and spatial axes, plus block sparse attention for efficient inference. It supports long video continuation without color drifting.
Yes. LongCat‑Video is MIT‑licensed and open source. You can explore the model on Hugging Face and the code on GitHub.
LongCat Video supports Text‑to‑Video, Image‑to‑Video, Video‑Continuation, and long‑video generation with robust text and image alignment.
The model weights are released under the MIT License. Review the license to ensure your LongCat Video use case complies with applicable terms and regulations.

Ready to Create with LongCat Video?

Join creators using LongCat Video and LongCat AI to produce minutes‑long, high‑quality videos with ease. Perfect for content, marketing, education, and storytelling—LongCat‑Video empowers your ideas.