What is Anthropic MCP?

The Model Context Protocol (MCP) is an open standard created by Anthropic for connecting AI models to external tools, data sources, and services. Announced in November 2024, MCP provides a universal interface that enables any AI agent to discover, connect to, and interact with any MCP-compatible server, creating an interoperable ecosystem for agentic AI applications.

MCP defines a client-server architecture where AI applications (clients) can dynamically discover available tools, execute them, and share context through a standardized protocol. The specification includes tool discovery, execution, resource access, and prompt templates. The November 2025 spec release introduced asynchronous operations, statelessness, server identity, and official extensions, bringing the protocol closer to production readiness.

In 2025, Anthropic donated MCP to the Agentic AI Foundation (AAIF), a directed fund under the Linux Foundation co-founded by Anthropic, Block, and OpenAI with support from Google, Microsoft, AWS, Cloudflare, and Bloomberg. The protocol now has an official community-driven Registry for discovering MCP servers, and major AI providers including OpenAI and Google DeepMind have adopted it. Claude's directory includes over 75 MCP connectors, and the ecosystem continues to grow rapidly.

Key features

Core capabilities this platform advertises.

Model Context Protocol specification
Official MCP SDKs (TypeScript, Python)
Server reference implementations
Standard tool interface
Open protocol

Strengths and tradeoffs

What this tool does well, and the limitations to keep in mind.

Pros

Open standard with broad industry adoption (OpenAI, Google, Microsoft)
Universal interface for connecting any AI model to any tool
Active community with official registry and 75+ connectors
Backed by Linux Foundation with major tech company support
Free and open-source with permissive Apache 2.0 license

Cons

Still evolving specification with breaking changes possible
Server implementation quality varies across community contributions
Requires development effort to build custom MCP servers
Security model for tool execution still maturing

Plans & pricing

What's included in each plan, and how the tiers compare.

MCP Protocol

Free

Open-source specification
Apache 2.0 license
TypeScript and Python SDKs
Community support

MCP Servers

Free

Open-source server implementations
Community-built connectors
Official registry

View official pricing page

Common use cases

Developers building AI tools and agents who want to follow the standardized MCP protocol

Standardized tool use for AI agents
Building interoperable MCP servers
Sharing context between AI applications
IDE and editor AI integrations
Agent-to-tool communication

Using Anthropic MCP with Respan

Respan supports MCP for extending AI agent capabilities with external tools and data sources. Teams can use MCP servers through Respan's gateway to add tool use capabilities to their LLM applications while maintaining full observability over tool calls, costs, and performance.

Monitor MCP tool calls alongside LLM requests
Track costs of tool-augmented AI workflows
Evaluate quality of tool-assisted LLM responses
Centralized logging of all MCP server interactions

Run Anthropic MCP in production with Respan

10KFree traces/mo

1,000+Models

5 minSetup

Respan lets you trace LLM and agent calls across any model or framework, A/B test prompts on production traffic, and route requests across 1,000+ models through one gateway.

Try Respan free