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.

A local go-livepeer orchestrator provides a real protocol-level test target for BYOC containers, AI pipelines, and payment flows. Running locally means faster iteration, no ETH cost, and full control over the orchestrator’s capability configuration.

​
Standalone Mode (No Ethereum)

For testing BYOC containers or AI pipelines without any Ethereum setup, start the orchestrator in standalone transcoder mode: 
livepeer \
  -transcoder \
  -orchAddr 127.0.0.1:8936 \
  -aiModels /path/to/aiModels.json \
  -metrics \
  -cliAddr 127.0.0.1:7936
Standalone transcoder mode processes jobs from a connected broadcaster without on-chain registration or payment validation. A local broadcaster configured with -orchAddr 127.0.0.1:8936 routes jobs directly to this orchestrator.

​
Full Orchestrator Mode

For testing the complete payment flow, run the orchestrator with an Ethereum account:

​
AI-Only Orchestrator

From go-livepeer v0.7.x, a single binary supports both transcoding and AI workloads. The -aiModels flag activates AI capabilities without any additional binary or separate process: 
livepeer \
  -orchestrator \
  -transcoder \
  -aiModels /path/to/aiModels.json \
  ...
This replaces the previous pattern of running a separate AI subnet binary. A local orchestrator in standalone transcoder mode is the fastest way to test BYOC containers and AI pipelines without any Ethereum setup.

BYOC Overview

BYOC container HTTP contract and aiModels.json registration fields.

PyTrickle Quickstart

Building a BYOC container with PyTrickle to test against a local orchestrator.

Local Gateway

Running a local broadcaster to route jobs to the local orchestrator.

Local Development Overview

Choosing the right local setup for your development scenario.
Last modified on May 19, 2026