Overview
LiveKit carries video as tracks. A robot publishes one track per camera view, and each operator subscribes to the views it needs. LiveKit encodes every frame, forwards it to each subscriber, and adapts the stream to the bandwidth available.
Publishing from a robot means capturing frames in your own code and pushing them into a video source. Your app controls the camera and the capture format. LiveKit handles encoding, simulcast, and delivery.
Video latency dictates how quickly an operator can react to what the robot sees. Encode in hardware where the platform provides it, and attach capture timestamps when you need to align frames with other sensor data.
In this section
Video tracks
Publish frames from a camera or other local source, and subscribe to frames from another participant.
Hardware encoder support
Use hardware video encoders to minimize encoding latency on supported platforms.
Timestamps and frame metadata
Attach timestamps and metadata to published frames to correlate video with sensor data.