MCP 101: Your first server
Welcome to Rinoova! This guide will show you how to take an existing REST API (described via OpenAPI) and turn it into a fully functional Model Context Protocol (MCP) server in minutes.
Step 1: Prepare your OpenAPI Spec
The core of the creation process is your OpenAPI specification. Rinoova supports JSON and YAML formats. You can upload the file directly or paste the public URL where your spec resides.
Step 2: Create a new Server and Version
From the dashboard, navigate to My Servers and click Create Server. Name your project. Then, you will be asked how to generate the tools. Select Import from OpenAPI URL or Upload OpenAPI File and provide your spec.
Step 3: Select Deployment Mode
Once the spec is imported, Rinoova will automatically discover all available endpoints and convert them into 'Tools' for AI agents. You can then choose your creation mode:
Fast Mode
Ideal if you have a public API. It merely requires the API_BASE_URL (e.g. https://api.yourdomain.com). Rinoova will automate the generation, build, and deploy all at once.
Manual Mode
Select this option if you need to configure complex environment variables, define server-to-server authentication (e.g. Bearer, API Key), or if you want to exclude/tweak specific tools before deploying.
Step 4: Configuration and Deployment
If you chose manual mode, you can review and refine the discovered tools. After that, you'll configure environment variables (the base URL of your API is mandatory, while other keys depend on necessary auth). Click on Deploy and you are done: Rinoova will compile the code into a Container and push it to the isolated cloud, returning an SSE endpoint ready to be used.
Step 5: Test and Connect
You can test the newly discovered or generated tools directly from the Cloud Dashboard via the Interactive Playground. If everything works as expected, the SSE Endpoint provided in the deployment is the key to connecting your fresh MCP server with the AI Agent of your choice (such as Cursor, Claude Desktop, or the Rinoova App).