concept

Model Context Protocol (MCP)

Model Context Protocol (MCP) is Anthropic's open standard for connecting AI assistants to external data sources and tools — letting any compliant client use any compliant server's capabilities.

MCP, introduced by Anthropic in late 2024 and broadly adopted by 2026, formalises a JSON-RPC interface between LLM clients (Claude Desktop, Cursor, Codex, Cline, custom agents) and servers that expose tools, resources, and prompt templates. The win is reusability: write one MCP server for filesystem, GitHub, Postgres, internal API, and every compliant client can use it. OpenAI, Anthropic, Google, Cursor, Continue, Zed all ship MCP-compatible clients. The pattern composes with function calling — the server defines the tools, the client routes through MCP, the model emits function calls.

When to use model context protocol (mcp)

Common mistakes

FAQ

What is model context protocol (mcp)?

Model Context Protocol (MCP) is Anthropic's open standard for connecting AI assistants to external data sources and tools — letting any compliant client use any compliant server's capabilities.

When should I use model context protocol (mcp)?

Exposing internal capabilities to multiple LLM clients. Avoiding vendor lock-in on tool integration. Distributing capabilities as reusable packages.

What are the most common mistakes with model context protocol (mcp)?

Confusing MCP with function calling — MCP is the protocol, function calling is the mechanism. Shipping MCP servers without auth or quotas — same security concerns as any RPC endpoint.

Sources

Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/model-context-protocol.md.