Model Context Protocol (MCP)
Definition
The Model Context Protocol (MCP) is an open standard, introduced by Anthropic, that defines a universal way for AI models to connect with external tools, data sources, and services. Instead of building custom integrations for each app, developers expose capabilities through MCP servers that any MCP-compatible client can use — standardizing how agents access context and take actions, much like a common port for AI connectivity.
How it works
MCP uses a client-server architecture. An MCP server wraps a data source or service — a database, a file system, a SaaS app, an internal API — and exposes its capabilities in a standardized way. An MCP client, embedded in an AI application or agent, connects to those servers and lets the model use them.
Servers can expose several kinds of capability:
-
Tools — actions the model can invoke, such as running a query or creating a record.
-
Resources — data the model can read, such as files or records.
-
Prompts — reusable templates that guide how a capability is used.
Because the interface is standardized, one MCP server works with any compatible client, and one client can talk to many servers. This replaces the combinatorial problem of building a custom connector for every model-and-tool pairing with a single shared protocol.
As an open standard, MCP has been adopted across a range of AI tools and providers, with a growing ecosystem of community and vendor-built servers.
Why it matters
Connecting AI models to real systems has historically meant bespoke, brittle integrations rebuilt for each model and each tool. MCP turns that into a write-once, reuse-everywhere pattern: expose a capability through an MCP server once, and any compatible agent can use it. This lowers the cost of building useful, grounded agents.
Standardization also improves portability and reduces lock-in. Because MCP is provider-neutral, the same tool servers can back agents built on different models, and switching providers does not require rewiring every integration.
The protocol is central to the agentic shift. Reliable tool and data access is what lets agents do real work, and MCP gives the ecosystem a common foundation for that access — alongside appropriate attention to permissions, authentication, and security, since connected tools can take real actions.
Frequently asked questions
Who created the Model Context Protocol?
MCP was introduced by Anthropic as an open standard. Being open, it is not limited to Anthropic's models — it has been adopted across a range of AI tools and providers, with a growing ecosystem of MCP servers.
How does MCP differ from regular function calling?
Function calling is how a model invokes a tool within a single application. MCP standardizes how tools and data sources are exposed in the first place, so the same server can be reused across many models and clients instead of each app defining tools its own way.
What can an MCP server expose to a model?
Typically tools (actions the model can invoke), resources (data the model can read), and prompts (reusable templates). A client connects to one or more servers and lets the model use these capabilities through a single, standardized interface.
Why is MCP important for AI agents?
Agents need reliable, reusable access to external tools and data to do real work. MCP provides a universal connector so agents can plug into many systems without custom integrations, improving portability and reducing lock-in across providers.
Anthropic
Anthropic is an AI safety and research company best known for the Claude family of AI models. It pioneered constitutional AI, a method for training models against a written set of principles, and created the Model Context Protocol (MCP), an open standard for connecting AI models to external tools and data sources.
Function calling / tool use
Function calling, also called tool use, is an AI capability that lets a model invoke external functions, APIs, and services to accomplish tasks beyond text generation. The developer describes available tools and their inputs; the model decides when to call one, emits structured arguments, receives the result, and uses it to continue. This connects language models to live data, code execution, and real-world actions.
Agentic workflows
Agentic workflows are AI architectures in which a model autonomously plans, calls tools, browses the web, executes code, and completes multi-step tasks with limited human input. Rather than producing a single answer, the system loops — observing results, revising its plan, and acting again — marking the shift from AI chat to AI work that carries out goals on a user's behalf.
AI agent
An AI agent is a software system that uses a large language model (typically GPT-4o, Claude 3.5 / 4 Sonnet, Gemini 2.5, or open-source equivalents) to plan, decide, and act over multiple steps to complete a goal — calling tools, retrieving data, and producing outputs without step-by-step human supervision. Agents are the working surface of agentic AI in 2026.
AI agent frameworks
AI agent frameworks are software libraries and platforms for building autonomous agents that plan, call tools, maintain memory, and complete multi-step tasks with language models. They provide the scaffolding — control loops, tool integrations, state management, and orchestration — so developers can assemble agents without rebuilding the agentic plumbing for every project. Examples include LangChain, LangGraph, CrewAI, and the OpenAI Agents SDK.
Claude
Claude is Anthropic's family of AI assistants, including the Sonnet and Opus models. It is known for long context windows, strong coding and reasoning, support for the Model Context Protocol, and a safety approach grounded in constitutional AI. Claude is used across consumer apps, developer APIs, and enterprise products.