The open network for
real-time AI video

Generate, transform, and interpret live video streams on a permissionless GPU network built for real-time video inference.

Use Cases

What you can build with real-time AI video

Build live generated worlds, AI avatars, real-time style transfer, and more — powered by Livepeer's open GPU network.

GENERATING
60 FPS
Frame 1,847 · 12ms

AI-Generated Worlds

Live environments inferred in real time from user inputs. Not pre-rendered, not pre-recorded. The world itself is the model's output.

INPUT
OUTPUT
Style: Anime · 22ms

AI Avatars

Video representations of AI agents that respond in real time, for tutoring, telepresence, and conversational interfaces.

LIVE
YOLOv8 · 8ms

Live Video Intelligence

Real-time analysis of live streams: sports broadcasts, security feeds, manufacturing lines. Processed during, not after.

IngestModelOutput
Pipeline: img2img · 3 stages
Active

Prompt-Driven Live Transformation

A creator goes live and the stream adapts: lighting, environment, visual style, based on a text prompt, in real time.

LIVE
1080p 60fps
720p 30fps
480p 30fps
360p 30fps
Bitrate: 4.2 Mbps · Latency: 85ms

Live Transcoding & Streaming

Adaptive bitrate transcoding across a global GPU network with sub-second latency.

The Network

Powered by independent GPU providers

Livepeer is a global network of independent GPU providers, builders, and engineers who run the network and shape its direction. Open-source. Permissionless.