Skip to main content

Introduction

Build realtime robotics apps with LiveKit for streaming, teleoperation, remote inference, and voice interactions.

Overview

LiveKit SDKs connect robots, vehicles, and other hardware devices for realtime multi-participant streaming of video, audio, and data from cameras and sensors, enabling remote operation, monitoring, and inference. Build voice interactions on top with the LiveKit Agents framework.

Use cases

LiveKit supports a range of realtime robotics apps, including teleoperation, realtime observability, remote inference, and voice interactions.

Supported platforms

LiveKit runs on a range of hardware platforms for robotics, from embedded Linux systems and Nvidia Jetson boards to low-cost ESP32 microcontrollers. It also integrates with common robotics frameworks like ROS and LeRobot through LiveKit Portal.

Hardware platforms

LiveKit SDKs run on the compute platforms most commonly used in robotics:

  • Embedded Linux systems: Single-board computers and custom Linux hardware.
  • Nvidia Jetson: GPU-accelerated modules with hardware-accelerated video encoding for onboard inference and streaming.
  • ESP32 microcontrollers: Low-cost, low-power devices for lightweight audio and data streaming.

Robotics frameworks

Connect an existing robotics stack to the cloud with LiveKit Portal:

  • ROS: Bridge a ROS graph to stream telemetry, run diagnostics, and control robots remotely. See LiveKit Portal for ROS.
  • LeRobot: Run end-to-end policies in the cloud on realtime streams from the robot. See LiveKit Portal.

In this section

Learn how to build realtime robotics apps with LiveKit.

Next steps

Explore runnable examples and starter apps for common robotics patterns:

Examples & starter apps

Runnable examples and starter apps for teleoperation, remote inference, and more.