Runway ML — Claude Code MCP server: features and setup
This MCP server connects Runway tools to Claude Code for media generation and task control. It needs a Runway API key and Node.js 20+.
claude mcp add --env RUNWAYML_API_SECRET=<RUNWAYML_API_SECRET> --env RUNWAY_REQUEST_TIMEOUT_MS=<RUNWAY_REQUEST_TIMEOUT_MS> --env RUNWAY_UPLOAD_TIMEOUT_MS=<RUNWAY_UPLOAD_TIMEOUT_MS> --env RUNWAY_ALLOWED_ROOT=<RUNWAY_ALLOWED_ROOT> --env RUNWAY_DOWNLOAD_ROOT=<RUNWAY_DOWNLOAD_ROOT> --transport stdio runway -- npx -y @mindstone/mcp-server-runwayWhat Runway ML does
An agent can ask Runway to make videos from text or still images, transform existing video, and produce images from prompts with optional references. Audio tools cover speech, sound effects, dubbing, and voice processing. Custom voice management is included.
Other tools track generation jobs, wait for completion, stop a job, and save an output to a local file. The server can also send media to Runway's temporary storage and report account balance or credit usage. It runs through npx.
For Claude Code, add it as an stdio MCP server and provide RUNWAYML_API_SECRET. Node.js 20+ and npm are required; without the Runway API key, the generation tools cannot access the service.
What it helps with
- turn a text prompt into a video
- animate a still image
- generate an image with optional reference images
- create spoken audio or a sound effect
- check a generation job and save its output locally
Example prompts for the agent
- Make a video from this description and check when the job finishes.
- Animate
poster.pngand save the resulting file locally. - Generate speech for this text, then check my Runway credit balance.
How to install Runway ML
Platform CLI
claude mcp add --env RUNWAYML_API_SECRET=<RUNWAYML_API_SECRET> --env RUNWAY_REQUEST_TIMEOUT_MS=<RUNWAY_REQUEST_TIMEOUT_MS> --env RUNWAY_UPLOAD_TIMEOUT_MS=<RUNWAY_UPLOAD_TIMEOUT_MS> --env RUNWAY_ALLOWED_ROOT=<RUNWAY_ALLOWED_ROOT> --env RUNWAY_DOWNLOAD_ROOT=<RUNWAY_DOWNLOAD_ROOT> --transport stdio runway -- npx -y @mindstone/mcp-server-runwayReplace <…> with your own values
Platform CLI
codex mcp add runway --env RUNWAYML_API_SECRET=<RUNWAYML_API_SECRET> --env RUNWAY_REQUEST_TIMEOUT_MS=<RUNWAY_REQUEST_TIMEOUT_MS> --env RUNWAY_UPLOAD_TIMEOUT_MS=<RUNWAY_UPLOAD_TIMEOUT_MS> --env RUNWAY_ALLOWED_ROOT=<RUNWAY_ALLOWED_ROOT> --env RUNWAY_DOWNLOAD_ROOT=<RUNWAY_DOWNLOAD_ROOT> -- npx -y @mindstone/mcp-server-runwayReplace <…> with your own values
Config file
{
"mcpServers": {
"runway": {
"command": "npx",
"args": [
"-y",
"@mindstone/mcp-server-runway"
],
"env": {
"RUNWAYML_API_SECRET": "<RUNWAYML_API_SECRET>",
"RUNWAY_REQUEST_TIMEOUT_MS": "<RUNWAY_REQUEST_TIMEOUT_MS>",
"RUNWAY_UPLOAD_TIMEOUT_MS": "<RUNWAY_UPLOAD_TIMEOUT_MS>",
"RUNWAY_ALLOWED_ROOT": "<RUNWAY_ALLOWED_ROOT>",
"RUNWAY_DOWNLOAD_ROOT": "<RUNWAY_DOWNLOAD_ROOT>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"runway": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@mindstone/mcp-server-runway"
],
"env": {
"RUNWAYML_API_SECRET": "<RUNWAYML_API_SECRET>",
"RUNWAY_REQUEST_TIMEOUT_MS": "<RUNWAY_REQUEST_TIMEOUT_MS>",
"RUNWAY_UPLOAD_TIMEOUT_MS": "<RUNWAY_UPLOAD_TIMEOUT_MS>",
"RUNWAY_ALLOWED_ROOT": "<RUNWAY_ALLOWED_ROOT>",
"RUNWAY_DOWNLOAD_ROOT": "<RUNWAY_DOWNLOAD_ROOT>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e RUNWAYML_API_SECRET=<RUNWAYML_API_SECRET> -e RUNWAY_REQUEST_TIMEOUT_MS=<RUNWAY_REQUEST_TIMEOUT_MS> -e RUNWAY_UPLOAD_TIMEOUT_MS=<RUNWAY_UPLOAD_TIMEOUT_MS> -e RUNWAY_ALLOWED_ROOT=<RUNWAY_ALLOWED_ROOT> -e RUNWAY_DOWNLOAD_ROOT=<RUNWAY_DOWNLOAD_ROOT> runway npx -y @mindstone/mcp-server-runwayReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Runway ML
How do I install Runway ML in Claude Code?
Run claude mcp add mcp-server-runway -e RUNWAYML_API_SECRET=<RUNWAYML_API_SECRET> -- npx -y @mindstone/mcp-server-runway. Replace the placeholder with your Runway API key.
What does Runway ML do?
It gives an agent access to Runway tools for video, image, and audio generation, voice management, and generation-job control.
What does the Runway ML MCP server require?
It requires Node.js 20+, npm, and a Runway API key supplied as RUNWAYML_API_SECRET.
