By Quick answer: What is MCP (Model Context Protocol)?
- What — an open standard connecting AI apps to tools and data.
- How — MCP servers expose tools and resources; MCP clients use them.
- Why — build once, work with many AI products.
- History — introduced by Anthropic (Nov 2024); donated to the Linux Foundation's Agentic AI Foundation (Dec 2025).
- Security — trust only reviewed servers; limit permissions.
Key takeaways
- MCP does for AI tools what USB did for peripherals: one common connector.
- Developers write one MCP server and reach many AI clients.
- Broad adoption across AI products made MCP the de facto standard.
- An MCP server is software with access to your systems — treat it with care.
- Approvals and least-privilege permissions keep MCP-powered agents safe.
Before MCP, connecting an AI assistant to your tools meant custom code for every pair: Slack for one assistant, Slack again for another, your database for a third. N tools times M AI apps — a mess. The Model Context Protocol fixes that with a common connector. Build an MCP server for your tool once, and any MCP-compatible AI client can use it.
How MCP works
MCP server
Exposes what a system can do — tools (actions), resources (data) and prompts (templates) — in a standard format.
MCP client
The AI application — an assistant, IDE or agent — that discovers and uses servers' capabilities.
The model
Decides when to call a tool based on the user's request; the client executes the call and returns results.
The official documentation describes the protocol, SDKs and reference servers3.
A quick example
- You ask your AI assistant: "Summarise last week's open bugs and draft a status update."
- The assistant sees an available issue-tracker MCP server.
- It calls the server's "search issues" tool for last week's open bugs.
- It reads the results and drafts the update.
- If configured, it asks your approval before posting anywhere.
Why it took off
Anthropic introduced MCP as an open standard in November 20241. It spread quickly across AI products; by December 2025 Anthropic reported more than 10,000 active public MCP servers and support in products including ChatGPT, Cursor, Gemini, Microsoft Copilot and Visual Studio Code — and donated MCP to the new Agentic AI Foundation under the Linux Foundation, co-founded with Block and OpenAI2. Neutral governance made it safer for everyone to build on.
MCP vs APIs
| Traditional API integration | MCP | |
|---|---|---|
| Built for | Software calling software | AI clients using tools |
| Integration effort | Per app, per AI product | One server, many clients |
| Discovery | Read documentation | Clients discover tools and their descriptions |
| Best when | You control both sides | You want broad AI compatibility |
Security basics
- Install trusted servers only — an MCP server can read or change your systems
- Least privilege — expose only the tools and data needed
- Require approval for actions that send, pay, delete or publish
- Treat retrieved content as untrusted — it can contain prompt injection4
- Log tool calls and review them
MCP is how many AI agents reach real systems. For product teams, see how to add AI to an existing app.
Key terms, defined
MCP — Model Context Protocol. MCP server — exposes tools, resources and prompts. MCP client — the AI app that uses servers. Tool — an action an AI can call. Resource — data an AI can read. Prompt injection — malicious instructions hidden in content.
Frequently asked questions
What does MCP stand for?
Model Context Protocol — an open protocol that standardises how AI applications connect to external tools, data sources and services.
Who created MCP?
Anthropic introduced the Model Context Protocol in November 2024 as an open standard. In December 2025, Anthropic donated MCP to the Agentic AI Foundation, a directed fund under the Linux Foundation co-founded with Block and OpenAI.
What is an MCP server?
A program that exposes capabilities — tools (actions), resources (data) and prompts — to AI clients in the MCP format. Examples include servers for file systems, databases, GitHub, design tools or a company's internal APIs.
Which apps support MCP?
According to Anthropic, MCP is supported in products including Claude, ChatGPT, Cursor, Gemini, Microsoft Copilot and Visual Studio Code, among many others.
Is MCP secure?
The protocol enables powerful access, so security depends on which servers you install and what they can do. Use trusted or reviewed servers, limit permissions, require approval for sensitive actions, and guard against prompt injection from untrusted content.
Do I need MCP to build AI features?
No. You can call tools directly from your own code. MCP is valuable when you want your tools usable across many AI clients, or want your AI client to use many existing tools without custom integrations.
Sources
- Anthropic — Introducing the Model Context Protocol (2024)
- Anthropic — Donating the Model Context Protocol and establishing the Agentic AI Foundation (2025)
- Model Context Protocol — Official documentation
- OWASP — Top 10 for Large Language Model Applications
- Checked October 23, 2026. Figures that change — fees, salaries, prices — are reviewed every six months; if something is out of date, tell us.
Where to go from here
- What is an AI agent? — what MCP connects
- How to add AI to an existing app — where MCP fits in product work
- Career guidesHow to get an AI agent developer job in 2027: workflows vs agents, tools and MCP, memory, guardrails and production operationsHow to get an AI agent developer job in 2027: workflows vs agents, tool and MCP design, memory, guardrails, oversight, pay and a portfolio plan.
- AI & vibe codingWhat is RAG? Retrieval-augmented generation explained simply — then properlyWhat RAG (retrieval-augmented generation) is, explained simply and then for builders: how retrieval works, embeddings, chunking, evaluation and when to use RAG.
- AI & vibe codingCursor vs Copilot vs Claude Code: which AI coding tool fits how you work?Cursor vs Claude Code vs GitHub Copilot compared on workflow, agent capabilities, codebase context, team controls and best use cases — with a decision guide.
- Career guidesAI skills for software developers in 2027: coding with assistants, building with LLM APIs, agents and MCP — while keeping code secure and correctAI skills developers need in 2027: coding with AI, reviewing AI code, building with LLM APIs and structured outputs, agents, MCP, security and evals.
- Tech stackWhat is an API? A plain-English explanation with a restaurant analogy and real app examplesWhat is an API? A plain-English explanation using a restaurant analogy, with real examples from UPI, maps and login, plus REST, keys, webhooks and security.
- AI & vibe codingWhat is an AI agent? A plain-English explanation, with examples you already useWhat an AI agent is, how it differs from a chatbot and a workflow, the loop that powers it, everyday examples and the risks.
Get the ideas, not the noise.
Sign up with Google and we’ll send you short UX psychology tips by email — why people tap, stall and leave, and what to do about it. Unsubscribe anytime. Prefer something to keep? Get the free Laws of UX guide as a PDF.
Make your AI feature something users trust.
A behavioral UX audit of your app or SaaS product: why users hesitate, stall and leave, and a prioritized fix list. Fixed price, 2–4 weeks.