Skip to main content
If you already have code, you can import a GitHub repository and deploy it directly as an MCP server.

​
Deployment steps

  1. Select Deploy a repository in the server list on the Dashboard
  2. Link your GitHub account and choose the repository to deploy (see GitHub integration for how to link)
  3. Arlet analyzes the repository and automatically determines the configuration needed to run it as an MCP server
  4. Review the analysis result and run the deployment. When it completes, the server is added to your MCP server list

​
After deploying

  • Auto deploy on push: pushing to the linked repository deploys your changes automatically
  • Environment variables: register secrets such as API keys in the server’s settings screen — no need to hardcode them
  • Logs: check deployment and tool-invocation logs on the Dashboard
Deployed servers are immediately usable from your connected AI clients through the gateway.