MCP Architecture

Understanding the three-layer architecture clarifies where auth, transport, and capability logic live — critical for debugging connection issues.

MCP architecture is a restaurant: host (dining room), client (waiter), server (kitchen) — each has a clear role in getting food (data/tools) to the customer (LLM).

Scenario

Cursor (host) spawns an MCP client that connects to a Postgres MCP server (server) over stdio, discovering query and schema tools at startup.

Solution

In Model Context Protocol, apply MCP Architecture to this scenario: Cursor (host) spawns an MCP client that connects to a Postgres MCP server (server) over stdio, discovering query and schema tools at startup. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to MCP Architecture (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where MCP Architecture happens in the code.

1# MCP Architecture — minimal example2from openai import OpenAI3
4client = OpenAI()  # create API client5
6# Ask the model to explain this topic7response = client.chat.completions.create(  # core API call for MCP Architecture8    model="gpt-4o-mini",9    messages=[10        {"role": "system", "content": "You explain mcp architecture clearly."},11        {"role": "user", "content": f"What is mcp architecture?"},12    ],13    temperature=0,14)15print(response.choices[0].message.content)  # show output for debugging

• Treating MCP Architecture as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for mcp architecture

Quick recap — the most important points from this module.