Back to Docs

Deployment Guides

Once you've generated your Model Context Protocol (MCP) server, the next step is making it available to Claude Desktop. We provide several deployment models to fit your organization's infrastructure.

Local Deployment (Testing)

The easiest way to get started is to run your generated MCP server locally via Node.js or Docker.

  1. Download the generated MCP server code from the dashboard.
  2. Install dependencies using npm install.
  3. Build the server using npm run build.
  4. Point your claude_desktop_config.json to the local Node process (e.g., node /path/to/build/index.js).

Production Deployment Strategies

Cloud Hosted

Deploy the server on AWS, Google Cloud, or Heroku as a long-running service. Claude Desktop can communicate with it via SSE (Server-Sent Events) transports.

Internal CI/CD (Pro/Business)

Automate the generation and deployment of the MCP server as part of your CI pipeline so it always stays in sync with your live OpenAPI spec.

Looking for a managed solution?

LegacyAI Business plans offer priority deployment support to help you get your MCP servers running securely within your enterprise.

Contact Sales