> ## Documentation Index
> Fetch the complete documentation index at: https://livepeer-4bd33421.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Actors Overview

> An overview of the Livepeer protocol actors.

Actors are the main participants in the protocol who make up the core functions of the network.

> [!NOTE]
>
> Explainer here...

## Quick Links

> [card] [Gateways](https://livepeer-4bd33421.mintlify.site/pages/01_about/livepeer-protocol/livepeer-actors/gateways.md)

> [card] [Orchestrators](https://livepeer-4bd33421.mintlify.site/pages/01_about/livepeer-protocol/livepeer-actors/orchestrators.md)

> [card] [Delegators](https://livepeer-4bd33421.mintlify.site/pages/01_about/livepeer-protocol/livepeer-actors/delegators.md)

> [card] [Builders & Users](https://livepeer-4bd33421.mintlify.site/pages/01_about/livepeer-protocol/livepeer-actors/end-users.md)

### Why This Architecture Matters

* Decentralized competition — Gateways and orchestrators can specialize
* Better developer UX — Apps talk only to Gateways
* Better performance — Routing optimizes for GPU availability & latency
* Alignment — Orchestrators focus on compute; Gateways focus on services

***

# Marketplace Architecture

The Livepeer Marketplace is an emerging ecosystem layer where Gateways and Orchestrators publicly advertise, price, and compete on real-time AI video services. This marketplace transforms Livepeer from a raw GPU network into a discoverable, composable, and economically aligned infrastructure layer.

***

## Why a Marketplace?

As Livepeer shifts toward real-time AI video (Daydream, ComfyStream, BYOC pipelines), the network needs:

* Service discovery
* Capability matching
* Transparent pricing
* Quality and performance competition
* A way for builders to choose providers intentionally

Gateways and Orchestrators participate jointly to make this possible.

***

## Marketplace Participants

### **Gateway Operators**

Gateways publish:

* Supported models (e.g., diffusion, ControlNet, IPAdapter)
* Supported pipelines (ComfyStream workflows, BYOC containers)
* Pricing (per frame, per second, per inference)
* Geographic/latency characteristics
* Performance metrics

Gateways serve as the “service storefront.”

***

### **Orchestrator Operators**

Orchestrators contribute:

* GPU capacity (A40, 4090, L40S, etc.)
* Model acceleration (TensorRT, Torch Compile)
* Latency and throughput guarantees
* Historical reliability/performance scores

Orchestrators are the “supply side” of compute.

***

## Marketplace Workflow

```mermaid theme={null}
flowchart TD

    A[Application<br/>Developer] --> B[Gateway Marketplace<br/>Capabilities, Models, Pricing]

    B -->|Select Gateway| C[Gateway<br/>Routing, Scheduling]

    C -->|Match GPU Supply| D[Orchestrator<br/>GPU Execution]

    D --> C --> A
```

```
```


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.