What is MCP?
The Model Context Protocol is an open standard for letting AI assistants discover and call external tools — like REST APIs, databases, and internal services — through a single, well-defined interface.
The 60-second definition
MCP — short for Model Context Protocol — was introduced by Anthropic in late 2024 and is now supported by Claude Desktop, Cursor, ChatGPT, Windsurf, Continue.dev, and OpenAI Agents.
An MCP server exposes three primitives:
- Tools — actions the model can invoke (call an API, query a database, send an email)
- Resources — read-only documents the model can attach to its context
- Prompts — reusable instructions the user can trigger
For most teams, tools are the entry point — and the only primitive easyMCP focuses on today.
How a tool call actually works
Tools are called over JSON-RPC. A typical request looks like this:
POST https://easymcp.eu/api/public/mcp/prj_x9k...
Content-Type: application/json
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "get_weather",
"arguments": { "city": "Berlin" }
},
"id": 1
}The MCP server runs the tool, returns the result, and the model uses it to compose its reply.
Why MCP matters
Before MCP, every AI assistant defined its own tool format. Building integrations meant writing the same logic three or four times. With MCP, one server works across every compatible assistant — your investment compounds.
Where easyMCP fits
Writing an MCP server by hand means setting up a Node or Python project, wiring up the SDK, defining JSON schemas, and hosting it somewhere stable. easyMCP collapses all of that into a visual builder. Build your first project →