What is an MCP server?
An MCP server is a service that exposes tools, resources, or prompts to AI assistants through the Model Context Protocol. Instead of each client inventing a custom integration, an MCP server gives compatible assistants a standard way to reach external systems.
How an MCP server works
The AI client connects to an MCP server, reads the server's available capabilities, asks the user to approve access where needed, and then calls tools or reads resources during a conversation or coding session.
Tools
Actions an assistant can call, such as querying a database, searching the web, sending a request, or controlling a browser.
Resources
Context the assistant can read, such as files, docs, tickets, project state, logs, or structured records.
Prompts
Reusable instructions or workflows that make a server easier to use from a supported AI client.
Common MCP server use cases
- Browser automation servers for Safari, Chrome, Playwright, and web testing.
- Database servers for Postgres, SQLite, MySQL, warehouses, and analytics tools.
- Developer servers for GitHub, filesystem access, CI, package managers, and issue trackers.
- Productivity servers for search, notes, calendars, documents, and internal APIs.
- UI and design servers for builders, component libraries, screenshots, and asset workflows.
How to choose an MCP server
Start with the job you want the assistant to perform. Then check whether the server supports your client, what permissions it needs, whether it runs locally or remotely, and whether the GitHub repository includes clear setup instructions.
Open MCP Server Hub Browse MCP servers