What is MCP? The USB-C port between AI and your website
A plain-language explanation of Anthropic's Model Context Protocol – and why this open standard matters for local businesses with online appointment booking.
2026-06-11
Remember the cable chaos?
There was a time when every device had its own plug. The phone used one, the camera another, the printer yet another. Bring the wrong adapter and you were stuck. Then came USB-C: a single port that fits almost everything. Laptop, headphones, phone – one cable for all of them.
The AI world had exactly the same problem for years. And the fix carries a clunky name: MCP, the Model Context Protocol.
So what actually is MCP?
MCP is an open standard introduced by the AI company Anthropic (the makers of the AI assistant Claude) on 25 November 2024. The purpose in one sentence: it lets an AI talk to your software and data in a simple, consistent way.
That sounds technical, but the core idea is simple. Until now, AI assistants were – in Anthropic's own words – "trapped behind information silos and legacy systems." They knew a lot, but couldn't look into your booking system, your calendar or your customer records. Every single connection required an expensive, one-off build by programmers.
MCP replaces fragmented one-off integrations with a single standard – the USB-C port between artificial intelligence and the software that keeps your business running.
Instead of building a separate "cable" for every program, there is now one universal port. That's why MCP has earned the nickname "USB-C for AI."
Why was MCP created?
Imagine having to call an electrician to hand-solder a custom cable every time you bought a new appliance. That's roughly how connecting AI to software used to work: laborious, costly and error-prone.
MCP solves three things at once:
- One standard instead of many islands: build an MCP connection once and any AI that understands the standard can use it.
- A secure two-way connection: the AI can not only read but also act – in a controlled way, with clear permissions.
- Open to everyone: MCP is not one company's secret, it's freely available.
Who is already using MCP?
A standard is only as good as its uptake – and here things moved fast. Companies like Block and Apollo were on board at launch, along with developer tools such as Zed, Replit, Codeium and Sourcegraph. Ready-made MCP connections existed from day one for the likes of Google Drive, Slack, GitHub and Postgres.
Crucially, the competition joined in too: OpenAI (the company behind ChatGPT) adopted MCP in March 2025, and Google DeepMind followed in April 2025. In December 2025, Anthropic handed the standard over to the Agentic AI Foundation under the Linux Foundation – so MCP no longer belongs to a single company, but to everyone. Just like USB-C.
What does this mean for your local business?
Here's where it gets concrete. Take a hair salon, a physio practice or a car repair shop with online booking on its website.
With an MCP connection, an AI assistant can talk directly to your booking system or customer database (CRM). A customer writes: "I need an appointment next week in the afternoon." The assistant can then:
- check free slots – live in the calendar, not from guesswork,
- suggest a suitable time and, if asked, book it straight away,
- recognise returning customers and answer the enquiry personally.
No manual back-and-forth, no enquiry missed at 10 p.m. The AI goes from being a friendly but deaf conversation partner to a genuine team member that's allowed to reach into your system – securely and under control.
How we make your website ready for the AI future
At fischerdev we build modern websites for local businesses – with an integrated appointment booking system. And we're already thinking one step ahead: websites that can plug into AI assistants via open standards like MCP, so availability is checked and appointments are booked automatically. Fast, mobile-optimised and prepared for a world where AI books alongside you.
Want to know what an AI-ready website with a booking system could look like for you? Let's talk – we'll explain it without the jargon.
Sources
- Anthropic: "Introducing the Model Context Protocol" — anthropic.com
- Wikipedia: "Model Context Protocol" — en.wikipedia.org
- Weights & Biases: "The Model Context Protocol (MCP) by Anthropic: Origins, functionality, and impact" — wandb.ai