Runway API

Runway API is an MCP server that connects AI assistants to Runway developer services for media generation and editing. Designed for creative developers, designers, and video producers, it enables LLMs to trigger text-to-image synthesis, image-to-video generation, video editing, video upscaling, and text-to-speech audio production directly through conversational interfaces. The server provides tools to query available models and inspect their specific parameters, ratios, and durations prior to execution. Payloads are validated locally against model constraints derived from OpenAPI specifications, preventing invalid requests from reaching the network. Users can monitor asynchronous rendering jobs, query task states, and cancel running operations. It can be deployed as a local extension using standard input and output or hosted as an HTTP remote server to serve web-based assistants like claude.ai, Cursor, and Zed, allowing users to authenticate requests using their Runway API keys.

Category: Design, Media & Creative

Tags: Generative AI, image generation, runway, video generation

Visit Runway API

How to install and configure Runway API

1. Clone the repository to your computer: bash git clone https://github.com/runwayml/runway-api-mcp-server cd runway-api-mcp-server 2. Install dependencies and compile the server: bash npm install npm run build 3. Add the server to your Claude Desktop configuration file (claude_desktop_config.json): json { "mcpServers": { "runway-api-mcp-server": { "command": "node", "args": [ "/ABSOLUTE_PATH_TO_YOUR_CLONED_REPO/build/index.js" ], "env": { "RUNWAYML_API_SECRET": "<YOUR_RUNWAY_API_KEY_HERE>", "MCP_TOOL_TIMEOUT": "1000000" } } } } 4. Alternatively, deploy as an HTTP server using npm run start:http or install directly via the Claude Desktop Extensions menu as an unpacked extension. Restart Claude to complete the setup.

What you can do with Runway API

• Generating synthetic video clips from reference images and prompt instructions using Runway models like Seedance. * Creating high-resolution images from detailed text prompts and existing image URLs directly inside assistant chats. * Upscaling low-resolution generated or existing videos to higher quality formats through automated API tasks. * Editing existing video media using prompt instructions and reference imagery via the Aleph editing model. * Producing spoken text-to-speech audio clips by specifying text prompts and custom voice selections.

Key facts

• https://github.com/runwayml/runway-api-mcp-server
• Design, Media & Creative
• Generative AI, image generation, runway, video generation

Part of MCP Servers

How do I install Runway API?

You can install the Runway API server by cloning its repository from GitHub, running npm install, and executing npm run build. Afterward, configure it in Claude Desktop by adding the node executable and build/index.js path to your claude_desktop_config.json file with your RUNWAYML_API_SECRET, or load it via Claude Desktop Settings as an unpacked extension.

What can Runway API do?

The Runway API MCP server lets language models call Runway creative tools. It supports listing supported models, generating images from text prompts and reference URLs, creating video clips from images, editing videos, upscaling video resolution, producing spoken audio files, and inspecting or canceling ongoing generation tasks.

Which MCP clients work with Runway API?

The server works locally with Claude Desktop via stdio. When launched as an HTTP remote service, it connects with web and desktop clients including claude.ai web, Cursor, Zed, and Windsurf using standard HTTP Bearer authentication.

How long do generated files remain available?

Images and assets generated through the Runway API remain available at their generated download links for only 24 hours. The server documentation recommends downloading output files immediately because they cannot be recovered once the link expires.