AI & LLMs
What is model Context Protocol (MCP)?
Model Context Protocol (MCP) is an open standard for connecting AI models to external tools and data sources, so any compatible model can use any compatible tool without custom integration work.
New to Bolt? Sign up free.
Technical definition
Model Context Protocol is a specification defining how AI applications communicate with external systems that expose tools, data, and prompts. An MCP server publishes its capabilities in a standard form. An MCP client, inside an AI application, discovers those capabilities and calls them on the model's behalf.
The protocol replaces point-to-point integration with a common interface. Without it, connecting five AI applications to five data sources requires 25 bespoke integrations. With it, each side implements the protocol once.
In plain terms
MCP is a standard plug. Before it, every AI tool needed a custom adapter for every service it wanted to reach. Now both sides build to one shape and connect.
Why it matters
MCP decides what an AI can reach, and reach is what separates a helpful assistant from a useful one.
An AI that can only read what you paste is limited by your typing. An AI connected to your project tracker, your database, and your documents can answer questions about your actual business.
For a small business, the near-term value is not building MCP servers. It is that the tools you already pay for increasingly expose one, so the AI you use can work with your real data instead of a description of it.
Access is also risk. An AI with a connection to a live system can read and change what that connection allows. Connect what the task needs and nothing more.
