Model Context Protocol

What is MCP server?

Definition: An MCP server is a program that exposes tools, resources or prompts to AI clients over the Model Context Protocol. Clients discover its tools via a tools/list call and invoke them through JSON-RPC.

A remote MCP server runs as an HTTPS service; a local one runs as a subprocess over stdio. Either way it answers an initialize handshake, declares its capabilities, and lists tools — each with a name, a description and a JSON-Schema input.

The quality bar that decides whether agents actually use a server: clear tool descriptions (agents pick tools by description), valid input schemas, low handshake latency, and a deliberate auth posture.

How do I know if my MCP server is healthy?

Run a real initialize + tools/list handshake and check protocol conformance, tool-description coverage, latency and TLS. A free scan surfaces all of it in seconds.

Put it into practice

Scan your MCP server →

Related terms