Skip to main content

​
Payments Flow

Gateways are consumers of compute services. They pay orchestrators to do work on behalf of their application customers. The payment chain:
  1. End users pay the gateway at the operator’s application-layer rate
  2. The gateway pays orchestrators via probabilistic micropayment tickets (ETH on Arbitrum)
  3. Orchestrators pay their workers (transcoders, AI runners) and redeem winning tickets on-chain
  4. The gateway keeps the difference between what customers pay and what orchestrators charge

​
Gateway Earnings

Gateway earnings come from business-layer margins. Gateways buy network services instead of earning protocol rewards.

​
Gateway Costs

Gateway costs are determined by the work you route and the orchestrator prices on the network.

​
Currency

Gateways pay orchestrators in ETH on Arbitrum. LPT remains the staking, governance, and reward token for orchestrators and delegators. ETH handles service payments while LPT handles protocol governance and staking. This keeps service costs denominated in a stable currency while LPT serves its governance function.

​
Gateway Operator Models

Four operator models show how gateways earn at the business layer:
Running your own gateway removes a third-party routing fee. If you are already building an application on Livepeer, self-hosting keeps that margin inside your stack.Who: App developers scaling past hosted APIs (Daydream, Livepeer.Cloud)Revenue model: Cost reduction. You keep the margin that a third-party gateway would otherwise charge.
Charge end users more than you pay orchestrators. The difference is your margin. This is the classic gateway business model.Who: Livepeer.Cloud SPE, LLM SPE, StreamplaceRevenue model: Per-request, per-minute, or subscription pricing to customers. Pay orchestrators at network rates.
Build a full platform on top of the gateway layer - API key management, multi-tenancy, billing, SLA monitoring, orchestrator tiering.Who: NaaP (Network as a Platform), enterprise integratorsRevenue model: SaaS pricing with usage-based billing. The gateway is the infrastructure layer of a larger product.
Run a gateway to ensure SLA guarantees on your orchestrators. Control orchestrator selection, failover, and pricing for your specific content.Who: Streaming platforms, media companies, live event producersRevenue model: The gateway supports the core business through stronger SLA control and workload routing.

​
Pricing and Fees

As a gateway operator, you set pricing at two levels:

​
Protocol-level costs (what you pay)

You control your maximum cost per job using CLI flags:
gateway cost limits
# Maximum you'll pay per pixel for transcoding
-maxPricePerUnit=1000

# Maximum you'll pay per AI capability/model
-maxPricePerCapability='{"capabilities_prices": [{"pipeline": "text-to-image", "model_id": "stabilityai/sd-turbo", "price_per_unit": 1000}]}'
These flags set an upper bound. The gateway will only route to orchestrators whose prices are at or below your maximum. Lower maximums mean fewer available orchestrators; higher maximums mean more options but higher costs.

​
Business Layer Pricing

Your fees to end users are set entirely at the application layer, outside the Livepeer protocol. Common approaches:
  1. Per-request pricing - charge per API call to your gateway
  2. Usage-based pricing - charge per minute of video or per AI generation
  3. Subscription models - monthly fees for access to your gateway services
application pricing example
// Your application code (not in Livepeer protocol)
func calculateUserPrice(requestType string, pixels int64) float64 {
    basePrice := getYourBusinessPrice(requestType)
    yourCost := getOrchestratorCost(pixels)
    profitMargin := 0.20 // 20% margin

    return basePrice + yourCost*profitMargin
}

​
Price Discovery

Use the CLI to discover current orchestrator market rates:
livepeer_cli price discovery
livepeer_cli
# Select "Set broadcast config" to see current market rates
# Then set your max prices accordingly
Your margin = what customers pay you - what you pay orchestrators.

​
Production Gateways

Streamplace operates a gateway providing video processing services to content creators. Creators pay Streamplace; Streamplace pays orchestrators. The margin is Streamplace’s revenue. Daydream operates a gateway providing AI video processing. Builders and creators pay Daydream; Daydream pays orchestrators. The margin is Daydream’s revenue. Both demonstrate the arbitrage model: the gateway sits between application demand and network supply, adding value through routing quality, developer experience, and service reliability.

Gateway Role

What gateways are and how the role has evolved.

Gateway Capabilities

Workload types, orchestrator selection, and session management.

Payment System

How probabilistic micropayments work.

Orchestrator Incentives

How the supply side (orchestrators) earns - for comparison.
Last modified on March 16, 2026