An open standard that lets AI models call external tools and services.
MCP — Model Context Protocol — is an open standard developed by Anthropic that defines how AI models communicate with external tools and services. Before MCP, every AI integration required custom code. MCP provides a single, consistent interface any AI tool can use.
At its core, MCP lets an AI model discover what tools are available, call those tools with structured inputs, and receive structured outputs. The model never needs to know how the tool works internally — just what it can ask it to do.
MCP supports two transport types: local (stdio, running on the same machine as the AI client) and remote (HTTP, running as a web service). Remote MCP servers are particularly powerful because they work across machines, behind SSH, and can be shared across multiple AI tools using the same API key.
The protocol is now supported by Claude Code, Cursor, Windsurf, GitHub Copilot (VS Code 1.102+), Gemini CLI, Zed, Continue, Cline, and many others — making it the de-facto standard for AI tool integrations.
In ClockMe
ClockMe runs a remote MCP server at clockme.co/api/mcp. Add your API key and the URL to any MCP-compatible AI tool and you gain 18 timer tools: start, stop, list entries, get reports, manage projects, and more. The same server works across all your AI tools — one API key, one setup, unified dashboard.
Try ClockMe free →MCP was created by Anthropic and open-sourced in November 2024. It has since been adopted by Cursor, Windsurf, GitHub Copilot, Gemini CLI, and dozens of other AI tools.
MCP is a lower-level protocol than most plugin systems. It defines a standard way for any AI to call any tool, rather than building tool-specific integrations. One MCP server can serve multiple AI clients without modification.
Not with ClockMe. ClockMe's MCP server runs remotely as a web service — you only need to add the URL and your API key to your AI tool's config. Nothing to install, nothing to run.