Decart

Founded by Dean Leitersdorf in 2024

Generate real-time interactive video worlds and edit live video streams with low latency

Pricing model

Demo

Rating

★ People love it

Time to value

Quick Setup (< 1 hour)

You can use Decart to build and run real-time generative video models and interactive simulations. It provides models like Lucy for live video editing, visual effects, character swaps, and real-time lip-syncing at thirty frames per second. You can also deploy Oasis to create interactive physical environments for robotics, autonomous driving, drones, and gaming simulation. Decart runs on an inference stack compatible with Nvidia GPUs, AWS Trainium, and Google TPUs, allowing developers to generate low-latency virtual worlds and transform live video feeds with sub-second response times.

What Decart does

Generate real-time video simulations from user keyboard inputsEdit live video streams at thirty frames per secondSynchronize digital avatar lip movements to real-time audioDeploy physical AI simulations for robotics and autonomous vehiclesAdjust video edit intensity using continuous slider controlsRun low-latency model inference across diverse hardware acceleratorsGenerates interactive open-world video simulations in real timeTransforms live video feeds at thirty frames per second with minimal latencyPerforms live lip-syncing and real-time character swapsProvides fine-grained video editing control using slider-based intensity adjustmentsSimulates photorealistic driving and physical environments for robotics trainingAccelerates inference execution across NVIDIA, AWS Trainium, and Google TPU chips