Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.livepeer.org/llms.txt

Use this file to discover all available pages before exploring further.

Livepeer applications combine hosted access, direct gateway access, BYOC pipelines, and frontend SDKs into one production path. This page gives application developers the decision frame for choosing the right access surface before committing to architecture. Use it to decide when a hosted API is enough, when a self-hosted gateway is warranted, and which build page to open next.

Running Your Own Gateway

NeedSelf-Hosted GatewayHosted API
Get Started FastNo, setup adds overheadYes, API key and go
Cost Savings At ScaleYes, direct Orchestrator settlement with no hosted-API markupNo, hosted provider margin sits on top of network price
Custom Orchestrator SelectionYes, pass any -orchAddr list or custom discovery endpointNo, provider controls routing
Data Stays Within Your InfrastructureYes, Gateway runs on your servers and requests stay in your stackNo, requests route through provider
Production Resilience And RedundancyYes, run multiple Gateways and control failover logicPartial, provider SLAs set the ceiling
Custom Auth Or Billing ModelYes, integrate your own user management, remote signer, or JWT layerNo, provider auth model only
Zero Infrastructure OverheadNo, you own the binary and the machineYes, nothing to run
The practical path is hosted access first, production validation second, and self-hosting once cost, routing, or data-control requirements justify owning the gateway. The main signals are monthly API spend, required Orchestrator selection, and infrastructure-bound inference paths.

Self-Hosting Requirements

This checklist states the operational requirements before you commit to self-hosting. The setup workflow lives in the Gateways tab.
Windows and macOS binaries for the AI Gateway are not currently available. Running a self-hosted AI Gateway requires Linux or Docker.
RequirementAI Gateway (Off-Chain)Video Gateway (On-Chain)
Operating SystemLinux or Docker on any hostLinux
ETH / On-Chain AccountNot requiredRequired: ETH account plus Arbitrum RPC URL
Staking / LPTNot requiredNot required for Gateway role
go-livepeer BinaryRequired: Linux binary or livepeer/go-livepeer:master Docker imageRequired: same binary
Orchestrator ListRequired: at least one -orchAddr endpoint to route toRequired: network discovery via on-chain signalling
Open PortPort 8937 (default) accessible from your appPort 8937 (default)
Time To First RequestAbout 15 minutes with Docker; longer for binary and configLonger: requires ETH account setup and on-chain registration
The AI Gateway path is designed for developers, not infrastructure operators. A single Docker command launches a functional Gateway. The on-chain video Gateway path is more involved and is primarily relevant to operators running the full Livepeer transcoding node.

The Two Gateway Types

TypeUse ForOn-ChainETH RequiredEntry Point
AI Gateway (Off-Chain)AI inference: text-to-image, LLM, ComfyStream, BYOCNoNoSet up an AI Gateway
Video Gateway (on-chain broadcaster)Video transcoding, HLS deliveryYesYesSet up a Video Gateway
The public Gateway at dream-gateway.livepeer.cloud and the Livepeer Studio AI API are both off-chain AI Gateway implementations of the same go-livepeer binary. When you self-host, you run that same binary yourself.
If you are ready to self-host, start with the local gateway setup. If you are not sure yet, the navigator helps you choose the right path for your use case.

Next steps

Set Up a Gateway

Full setup guide for self-hosted AI and video Gateways in the Gateways tab.

Back to the AI API

Not ready to self-host yet. Return to the hosted API quickstart.

What is a Gateway?

Understand how Gateways work before deciding whether to run one.
Last modified on May 19, 2026