In this hands-on workshop, we will build an MCP server and connect it to an AI agent (such as Cursor or Postman) to simulate real-world tool orchestration. We will move step by step through the following phases:

Phase 1 – Foundation:
Understand why standardized engineering with FastAPI turns AI from demo to production system. Validate your setup (Git, Python, VSCode, Postman/Cursor). Explore the template-mcp-server repo and its architecture.

Phase 2 – Zero to Local:
Clone the repo, configure environment files, and run the base template. Review project structure, request flow, and where core logic resides.

Phase 3 – Build a Custom Tool:
Implement a new MCP tool step-by-step. Define its purpose, inputs, and execution logic. Run locally and confirm it’s discoverable and functional via testing.

Phase 4 – Integration & Orchestration:
Connect your MCP server to an agent (e.g., Cursor). Observe live tool invocation. Finally, containerise the app for scalable, production-ready deployment.