← Learn

Model Context Protocol (MCP)

An open standard for connecting AI systems to tools and data sources through a shared protocol, so each integration does not need a custom connector.

Learn

When to use it

Use Model Context Protocol (MCP) when you need to connect AI systems to multiple tools and data sources without creating custom connectors for each one. MCP unifies these connections under a shared protocol, enabling seamless integration with things like databases, APIs, and analytics platforms.

Quick example

In Amazon Bedrock, your AI agent needs to interact with different data sources and external tools for prospecting tasks. You configure MCP within Bedrock to handle these interactions through a single protocol, eliminating the need for multiple custom connectors. Bedrock acts as the client, while MCP standardizes the connection to various resources.

prompt → MCP → databases/APIs → process data → stop

Ecosystem

MCP operates alongside AI models and tools, providing a standardized connection layer. It simplifies the integration process, allowing models to access external resources efficiently.

        ┌─ databases/APIs ─┐
model →│   MCP   │→ process data
        └─ tools ────────┘

Misconceptions

MisconceptionRebuttal
MCP changes model behaviorMCP only standardizes connections, not model logic
Any protocol can replace MCPMCP is specifically designed for AI tool integration
MCP is a standalone toolMCP is a protocol, not a separate tool

Trade-offs

  • Integration simplicity — requires protocol compliance
  • Scalability — may add overhead for small-scale systems
  • Standardization — limits flexibility in custom connections

Seen in