Skip to main content

​
Deployment

A deployment is the complete configuration of a Livepeer gateway, defined by three independent choices: operational mode, setup type, and node type. This guide will walk through deployment of a gateway using go-livepeer and cover both on-chain and off-chain modes using Docker, Windows & Linux OS environments with node configurations for video, AI, and dual workloads.

​
Setup Type: go-livepeer

The setup type determines what software you run and how you install it. The reference setup type is running the go-livepeer gateway software, whichcan be installed via source binary or Docker and is the recommended setup for most users as it provides the most features, stability, and compatibility with the Livepeer network.For alternative setup types see the Gateway Setup Options page, which covers using the Livepeer SDK, running a gateway with a GWID, and using hosted gateway services.

​
Operating System: Docker, Linux, Windows

You can run a Livepeer gateway on either Linux or Windows, but Linux is recommended for production deployments due to better performance, stability, and compatibility with GPU drivers and containerization tools. Windows is suitable for development and testing.

​
Operational Modes

The operational mode determines how your gateway integrates with the Livepeer protocol. You can run a gateway both independently: Off-chain or connected to the blockchain-based Livepeer network: On-chain. Off-chain
  • Run a gateway that operates independently of the blockchain. This is ideal for testing, development, and private deployments. Off-chain gateways do not interact with the blockchain for orchestrator discovery or payment settlement - they rely on manual configuration and external payment mechanisms.
On-chain
  • Run a gateway that is fully connected to the blockchain-based Livepeer network (on Arbitrum). On-chain gateways interact with the blockchain for orchestrator discovery, payment settlement, and ETH management.

​
Node Types

The node type determines the workloads your gateway can route. You can run a Gateway in the following node configurations:
  • Video Only -> traditional transcoding services
  • AI Only -> AI inference services
  • Dual: AI & Video -> both video transcoding and AI inference services

​
Gateway Operator Journey

​
Gateway Architecture

Alternative setup options for running a gateway using the Livepeer SDK, GWID, or hosted services.

Looking for hands-on walkthroughs to run your first gateway and test it end to end?

Looking for information on how gateways earn fees for services?

Want to learn more about running AI workloads on Livepeer Gateways?
Last modified on April 7, 2026