Mirage LSD
Real-time video-to-video AI that applies user-prompted, dynamic style transfer
See for yourself
Finally, Infinite Generation
The auto-regressive nature of the model leads it to accumulate errors, effectively limiting the length of its outputs. To enable infinite auto-regressive generation:


The Only Live-Stream
Diffusion AI Video Model
Mirage LSD V2 takes realtime video to a new level. completely transform and restyle — without losing a detail.
Zero Latency
<40ms per frame, a realtime video experience like never before
Open ended prompts
Describe an idea, see it live. chose from curated suggestions or prompt your own.
Detail preservation
Transforming entire scenes without losing a detail, while maintaining characters recognizable.
A New Creative Layer Across Every Industry
Live Events
Use real-time video AI to transform audiences, stages, and screens with live, dynamic visuals that react instantly.
Gaming
Bring real-time transformations into gameplay and streams, updating environments and visuals as players interact.
Streaming
Enhance live broadcasts with real-time video changes, enabling instant styling, overlays, and interactive visual updates.
XR
Layer real-time video AI into XR environments to create immersive, responsive experiences that adapt to movement and space.
Production
Real-time generative overlays and look-development on set
Advertising
Enable real-time, in-stream ad personalization with dynamic visual changes that adapt to context, audience, and placement.
const formData = new FormData();
formData.append("video", fs.createReadStream("/path/to/input.mp4"));
formData.append(
"prompt",
`Replace the human with a bipedal tiger-like humanoid.It has orange-and-black striped fur, a powerful upright stance,muscular arms, and a focused feline gaze`
);
const response = await fetch("https://cdn.api.decart.ai/vid2vid/process", {
method: "POST",
headers: {
"X-API-KEY": process.env.DECART_API_KEY,
Ready to try
Mirage LSD?
Generate infinite, temporally-coherent video in real time with just a single API call.
FAQ
Decart AI is a real-time generative video platform that lets developers create live, reactive video experiences. You can build interactive live streams, AR overlays, marketing videos, or any application that requires fast, low-latency video generation.
Our API is optimized for ultra-low latency, with real-time responses under 100ms. This ensures smooth performance for live applications, interactive installations, and instant previews.
Developers can integrate Decart AI using our SDKs, plugins, and API keys. We provide code snippets, documentation, and example projects to get started quickly, whether you’re using Python, JavaScript, or other popular languages.
Decart supports multiple generative models including Real-Time AI Video Editing and Restyling, Video-to-Video, Image-to-Image, and Text-to-Video generation. You can test models in our Playground before integrating them into your app.
We offer a free API tokens for developers to test and explore real-time video generation. Pricing scales with usage and is designed to be flexible for builders, startups, and enterprise applications. Contact us for enterprise-tier solutions.
If you need help or have questions, you can contact our support team at contact@decart.ai Our team is here to assist with any issues related to your account, generations, or pricing