# Relay

> Relay is an MCP server for design and media. Run image, video, audio and LLM models on Relay's GPU network from your agent. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Platforms:** API, CLI
- **Website:** https://docs.relaygpu.com/sdk/mcp
- **Repository:** https://github.com/opengpu-network/relaygpu-mcp
- **MCP Registry name:** com.relaygpu/mcp
- **Categories:** [design and media](https://ai-powered.com/categories/design-and-media), [AI models and LLM tooling](https://ai-powered.com/categories/ai-and-llm)
- **Listing updated:** 2026-11-18

## About

Tools here are generic and take the model name as a string, so a model added to Relay works without a new release. An agent can find a model, read its input schema, run it, poll long-running tasks, upload files, run workflows and search Relay's docs.

There are two ways to connect. The hosted Streamable HTTP server at mcp.relaygpu.com/mcp needs nothing installed and takes your key in an X-API-Key header or as a bearer token. The local stdio package @relaygpu/mcp runs with npx on Node 20 or newer and reads RELAY_API_KEY from its environment. Per the docs, the hosted server reads the key from each request and uses it for that request only.

Claude Desktop does not yet get remote OAuth from Relay, so it uses a one-click .mcpb bundle that runs the local server. The get_credits and get_usage tools need a dashboard JWT or a custom-tier superkey rather than a plain key.

## Features

- **Model-agnostic tools:** Tools accept the model as a string, so new Relay models need no server update.
- **Task polling:** Start long image or video jobs and poll them until finished.
- **File upload:** Upload inputs for models directly from the agent.
- **Workflow runs:** Run Relay workflows through the server.
- **Docs search:** Search Relay's documentation from the chat.
- **Hosted or local:** Use the hosted endpoint or run the npm package over stdio.

## Good for

- Generate an image or video from a coding agent
- Check which Relay models exist and read their input schema
- Run a long task and poll until it completes
- Look up Relay API docs without leaving the editor

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http mcp https://mcp.relaygpu.com/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://mcp.relaygpu.com/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "mcp": {
      "url": "https://mcp.relaygpu.com/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "mcp": {
      "type": "http",
      "url": "https://mcp.relaygpu.com/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "mcp": {
      "serverUrl": "https://mcp.relaygpu.com/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "mcp": {
      "type": "http",
      "url": "https://mcp.relaygpu.com/mcp"
    }
  }
}
```

## Tools

- `get_credits`: Reads account credits (needs a dashboard JWT or superkey)
- `get_usage`: Reads account usage (needs a dashboard JWT or superkey)

## FAQ

### Do I need an API key?

Yes, a Relay key starting relay_sk_ from dashboard.relaygpu.com/keys.

### Which clients are covered?

Claude Code, Codex, Cursor and VS Code use the hosted server; Claude Desktop uses a local bundle.

### Does the server store my key?

The docs say no: the hosted server uses it per request, and the local one reads it from the environment.

### What Node version does the local server need?

Node 20 or newer.

### How do I add Relay to Claude Code?

Run this command in your terminal: claude mcp add --transport http mcp https://mcp.relaygpu.com/mcp

### What are the best alternatives to Relay?

The closest alternatives to Relay are HyperFrames by HeyGen, headroom, omniget, next-ai-draw-io and mcp-for-blender.

## Where this comes from

- Description: written from the project's README and the project's website.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [HyperFrames by HeyGen](https://ai-powered.com/mcp/hyperframes-by-heygen): HyperFrames by HeyGen is a freemium MCP server for design and media.
- [headroom](https://ai-powered.com/mcp/headroom-3): headroom is an open-source MCP server for ai models and llm tooling.
- [omniget](https://ai-powered.com/mcp/omniget): omniget is an open-source MCP server for ai models and llm tooling.
- [next-ai-draw-io](https://ai-powered.com/mcp/next-ai-draw-io-2): next-ai-draw-io is an open-source MCP server for ai models and llm tooling.
- [mcp-for-blender](https://ai-powered.com/mcp/mcp-for-blender): mcp-for-blender is an open-source MCP server for design and media.
- [ffmpeg-skill](https://ai-powered.com/mcp/ffmpeg-skill): ffmpeg-skill is an open-source MCP server for design and media.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [unity-mcp](https://ai-powered.com/mcp/unity-mcp-2): unity-mcp is an open-source MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/relay-3
