Get up and running with your first MCP server in just a few steps. This guide walks you through signing in, deploying an MCP server, testing it in the Playground, and viewing traces.

​
Step 1: Sign In or Create an Account

Go to platform.contexaai.com and sign in using your credentials. Don’t have an account? Click Sign Up to register in seconds.

​
Step 2: Deploy Your MCP Server

Choose one of the following ways to deploy: 🔍 Option 1: Use a Curated Server Navigate to the MCP Directory. Browse verified servers contributed by the verified developers and organizations. Click Deploy on any server to set it up instantly. 🔗 Option 2: Bring Your Own Server Go to Directory > Add Server > via GitHub. Deploy via Github Enter public GitHub repo URL containing your MCP code. Configure deployment settings and click Deploy. 📄 Option 3: Create from OpenAPI Spec Navigate to Directory > Add Server > via OpenAPI. Deploy via OpenAPI Upload your OpenAPI 3.0 spec file or paste the spec in the editor. Name your server and click Deploy.

​
Step 3: Test in the Playground

Once your MCP server is deployed:
  1. Go to the Playground section.
  2. Select your MCP server from the dropdown list.
  3. Choose an LLM model (e.g., GPT-4, GPT-4o, GPT-4o-mini).
  4. Send a request and see live responses from your MCP server.
  5. Optionally, compare outputs across different models to evaluate performance.

​
Step 4: View Traces & Logs

  • Navigate to Traces.
  • Platform Logs: See all tool calls with timestamps, inputs, outputs, and status.
  • Server Session Logs: View grouped logs by session for detailed analysis.
  • Use filters to drill down into specific tools or calls.

​
✅ You’re All Set!

You’ve now:
  • ✅ Signed in
  • ✅ Deployed your first MCP server
  • ✅ Tested it with real models
  • ✅ Viewed logs and metrics
Start building intelligent, composable tools with the power of Model Context Protocols and LLMs.