This index covers operator-side setup. If you are a developer looking to use a gateway via the API or SDK, see the Developers tab.
Getting started
Official Gateway Setup Guide
The canonical operator quickstart in these docs. Binary or Docker, gateway flags, verify the node is running. Start here.
Livepeer Gateway Path
Decision matrix: video vs AI gateway, on-chain vs off-chain, hosted service vs self-hosted. Choose your entry point before touching any config.
go-livepeer Releases (GitHub)
Official binary releases for Linux (amd64 and arm64). Always download a tagged release for production use.
Why Run a Gateway
Business case for self-hosting: cost savings at scale, full orchestrator control, lower latency, and inference fee capture.
Off-chain AI gateway setup
These guides cover the recommended path for AI workloads: the off-chain gateway with a remote signer, introduced in Q4 2025 via PRs #3791 and #3822.Off-chain mode removes all Ethereum dependencies from the gateway. No ETH, no Arbitrum RPC, no local signing key — all of that is handled by the remote signer. This is the correct starting point for all new AI gateway deployments.
AI Gateway Quickstart (Off-chain)
Minimal command to run an off-chain AI gateway with Docker, pointed at a community remote signer and a local or public orchestrator.
Remote Signers: Design and Setup
Technical explanation of what a remote signer does, the signing protocol (getOrchInfoSig → signTicket), stateless design, and how to point your gateway at one.
Elite Encoder Remote Signer (Community-hosted)
Community-hosted remote signer with free ETH for testing. Confirmed operational by Elite Encoder in #local-gateways Discord.
Orchestrator Discovery for Off-chain Gateways
Three discovery methods: explicit orchAddr list, discovery webhook URL, or the remote signer GetOrchestrators endpoint. Python gateway SDK examples included.
Remote Signers Design Doc (PR #3822)
Canonical technical specification for remote signing by j0sh. PM bookkeeping, state design, alternatives considered, and security implications.
AI Configuration Reference
AI-specific flags:
-aiWorker, -orchAddr, -remoteSignerAddr, ports, and startup log verification.BYOC — Bring Your Own Container
BYOC lets you attach any Docker container as a compute pipeline on the Livepeer network. It is the primary path for running custom AI workloads — including CPU-only pipelines — without a GPU.BYOC Smoke-Test Tutorial: CPU Gateway + Orchestrator (Off-chain)
Step-by-step: run a gateway and orchestrator on one machine using a CPU Docker container, test end-to-end with off-chain payments via the community remote signer, then take it to production on-chain.
Creating a Custom Live Pipeline (ai-runner)
Official guide: implement the Pipeline interface, define custom params, build a Docker image extending
livepeer/ai-runner:live-base. Uses scope-runner as a reference.Scope Runner (Reference BYOC Implementation)
Production reference implementation of a custom BYOC pipeline. Use as a template when building your own container.
BYOC Payments (PR #3866)
BYOC remote signer payments PR by Elite Encoder. Token refresh, remote signer changes for external capabilities, and how BYOC discovery differs from LV2V (no gRPC for OrchestratorInfo).
Python Gateway SDK with BYOC
livepeer-python-gateway supports BYOC session management, orchestrator discovery, and remote signer payments. The
examples/ folder is the best starting point.ComfyUI BYOC Integration (Elite Encoder)
BYOC implementation using livepeer-python-gateway as a ComfyUI node backend. Tested stable for multi-minute streams with LV2V payments and remote signer from main.
Video gateway (on-chain transcoding)
Run a Video Gateway
Full setup for a video transcoding gateway: RTMP input on port 1935, on-chain ETH key, Arbitrum RPC, and orchestrator connection.
On-chain Requirements
ETH deposit and reserve requirements, Arbitrum RPC setup, wallet management, and the
livepeer_cli commands to deposit and activate.Video Configuration
Video-specific flags:
-broadcaster, RTMP port, transcode profiles, segment size, and webhook auth config.Advanced Transcoding Config
Output profiles, per-capability pricing, and
setbroadcastconfig CLI commands for production transcoding setups.Bridge LPT to Arbitrum
How to move LPT from Ethereum mainnet to Arbitrum One for on-chain registration and staking.
Livepeer Explorer
Monitor your gateway’s activity, connected orchestrators, and payment ticket activity on-chain.
Dual gateway (video and AI on the same node)
Dual Configuration Guide
Run video transcoding and AI inference from the same go-livepeer instance. Port assignments, flag combinations, and pitfalls to avoid.
Gateway Architecture Overview
System diagram covering on-chain gateway, off-chain gateway, and dual-mode operation. Updated for remote signers and the clearinghouse model.
Payments and economics
How Payments Work
PM mechanics: ticket face value, per-pixel pricing for AI inference, per-segment pricing for video, deposit/reserve lifecycle.
Pricing Configuration
Set
-maxPricePerUnit and -maxPricePerCapability, format the AI pricing JSON, and understand price negotiation between gateway and orchestrators.Payment Clearinghouse
What a clearinghouse is: remote signer + user management + auth + accounting. Current status (active development) and NaaP dashboard as the reference implementation.
NaaP Dashboard (JWT Auth Demo)
Network as a Platform dashboard. Implements JWT issuer service (SIWE → user-scoped JWT) and Developer API Keys. Reference for building user auth on top of a remote signer.
Arbitrum Exchanges Reference
Exchanges where you can acquire ETH on Arbitrum One for gateway PM deposits.
Orchestrator discovery
Connect with Orchestrator Offerings
How an on-chain gateway discovers and selects orchestrators: stake-weighted for video, capability-based for AI.
Discover Orchestrator Offerings
Browsing available orchestrators by capability, model support, and pricing via the Livepeer Marketplace.
select_orchestrator.py Example (SDK)
Code example: discovery via the remote signer’s GetOrchestrators endpoint, filtered by model ID and capability. First-response-wins selection strategy.
Elite Encoder: Orch List Endpoint Proposal
A simple HTTP endpoint for off-chain gateways to retrieve a list of on-chain orchestrators without a full Ethereum dependency. Proposed by Elite Encoder in #local-gateways.
Monitoring and operations
Monitor and Optimise
Key metrics, ETH balance monitoring, ticket redemption rates, and log patterns indicating payment or session failures.
Gateway Introspection with Loki
Query gateway logs via the Loki API at
loki.livepeer.report. Pre-built queries for session starts, payment failures, and orchestrator errors.Gateway Middleware
Add auth, rate limiting, and request routing in front of your gateway. Patterns for the Gateway-as-a-Service provider persona using go-livepeer webhook auth hooks.
Production Log Monitoring Guide (Forum)
Forum guide on structured logging, alerting, and pattern matching for production gateway nodes. Search “Production Log Monitoring” on the forum.
tools.livepeer.cloud — Gateway Tools
Community tool suite for network operators: stream tester, orchestrator explorer, performance dashboards. Run by the Livepeer.Cloud SPE.
Community SDKs and alternative gateway implementations
livepeer-python-gateway (j0sh)
Reference Python gateway SDK. Implements LV2V payments, remote signer integration, BYOC support, orchestrator discovery, and session management. Actively maintained by j0sh (Livepeer).
OrchestratorSession Class (Elite Encoder PR)
Combines discovery, selection, payment refresh, and stream management in a single SDK entry point. Spec documented in Elite Encoder’s gist.
j0sh Ecosystem Notes Gist
j0sh’s working notes on the gateway ecosystem: SDK design, clearinghouse thinking, discovery architecture, and open problems. Updated periodically.
Elite Encoder OrchestratorSession Spec
Spec for the OrchestratorSession class tying payment refresh, publish, and subscription together as one managed session. The foundation of the BYOC SDK work.
Gateway-as-a-Service (Provider persona)
Operator Support & Programmes
Revenue models: inference fees, SPE grants, Foundation programmes, and the NaaP approach to building a gateway business on top of the Livepeer network.
Livepeer.Cloud — How to Run a Gateway Node
Livepeer.Cloud SPE’s own operator guide. Real-world operational detail from an active multi-region gateway provider.
NaaP: Network as a Platform
Dashboard and tooling for Gateway-as-a-Service operators. API key management, JWT auth, Developer API Keys, and remote signer plugin integration.
SPE Grants and Treasury Proposals
Foundation grants for gateway operators. Existing examples include the Cloud SPE and LLM SPE. Apply via the grants programme or governance forum.
Blog posts and deep dives
A Real-Time Update to the Livepeer Network Vision (Nov 2025)
Clearest statement of the gateway’s role: routing AI inference, 72% of fees from AI, the path to a diverse gateway ecosystem.
Introducing Livepeer Cascade (Feb 2025)
Strategic vision. Gateways are the distribution layer for real-time AI video. Contextualises the off-chain and remote signer investment.
Building Real-Time AI Video Effects with ComfyStream (Feb 2025)
ComfyStream as a backend inference engine via a Livepeer gateway → orchestrator pipeline. Architecture, frame rates, and live demo.
ComfyUI and Real-Time Video AI Processing (Jan 2025)
DAG architecture, pipeline design, and real-time use cases — from the gateway operator’s perspective.
Builder Story: StreamDiffusionTD + Daydream API (Sep 2025)
Gateway + BYOC-style pipeline powering a live generative video performance using TouchDesigner + StreamDiffusion.
Introducing Daydream Beta (Sep 2025)
Daydream is a hosted gateway and API layer on top of the Livepeer network. Useful reading to understand what a hosted gateway provides before deciding whether to self-host.
Video walkthroughs
Livepeer Gateway Node Setup (YouTube — Titan Node)
Community video walkthrough of the full gateway setup process. Covers binary install, flags, and on-chain registration.
ComfyStream Live AI Demo
Video demo of ComfyStream processing live video via a gateway → orchestrator pipeline with real-time AI effects.
Livepeer Dev Office Hours (bi-weekly)
Bi-weekly video calls hosted by the Foundation. ComfyUI demos, coordinated gateway development, ecosystem priority updates. Calendar link in the Dev Hub.
Missing a guide? Open an issue at github.com/livepeer/docs/issues or post in #lounge on Discord.