vMira

What is MCP server?

An MCP server is a process that exposes tools, resources, and prompts to LLM clients over the Model Context Protocol — an open standard (introduced by Anthropic, widely adopted across the industry by 2026) for connecting AI assistants to external systems without bespoke per-vendor integrations.

Also known as: Model Context Protocol server, MCP, MCP integration

The problem MCP solves

Before MCP, every tool integration was bespoke per LLM provider: write OpenAI function definitions for ChatGPT, Anthropic tool definitions for Claude, custom schemas for everyone else. MCP replaces this with one server interface that any compliant LLM client can connect to — the same way HTTP replaced bespoke protocols between web clients and servers.

What an MCP server exposes

Three things: (1) Tools — functions the LLM can invoke (read a file, query a database, call an API). (2) Resources — content the LLM can read (file contents, search results, sensor data). (3) Prompts — reusable prompt templates the user or LLM can invoke. The MCP client negotiates available capabilities at connection time.

Where MCP servers run

Locally (stdio transport) — common for filesystem, git, IDE integrations. Remotely (HTTP/SSE or WebSocket transport) — common for SaaS integrations (Linear, Slack, Notion, Stripe). The protocol is transport-agnostic; what matters is the JSON-RPC message shape.

Adoption in 2026

Claude Code, Cursor, Continue, and most major IDE-AI integrations support MCP natively. OpenAI added MCP support in 2025. There are hundreds of community-maintained MCP servers for popular SaaS tools. For developers, writing an MCP server is a more durable investment than writing per-vendor integrations because the same server works across every compliant client.

Last updated · First published

Related terms

Tool use

Tool use is the LLM capability where, instead of producing a normal text response, the model emits a structured request to invoke an external function or API. The host application executes the function, returns the result to the model, and the model continues the conversation using that result.

AI agent

An AI agent is an LLM-driven system that can take multiple actions toward a goal — invoking tools, reading results, deciding next steps, and iterating — rather than producing a single chat response. The defining feature is autonomy: the agent decides what to do next based on what it has seen so far.

OpenAI-compatible API

An OpenAI-compatible API exposes endpoints and request/response shapes that match OpenAI's REST API (chat/completions, embeddings, models). Pointing the OpenAI SDK at the alternative provider's base URL with a new API key works without code changes.

Prompt engineering

Prompt engineering is the practice of designing the text input given to an LLM to maximize the quality, reliability, and specificity of the output. It covers everything from a one-line instruction to a multi-thousand-token system prompt with examples, constraints, and structured output schemas.

Try MCP server in vMira

Open the workspace and explore — no credit card required.

Open vMira