An MCP server connects an AI application, such as ChatGPT or Claude, to external tools, information or services. It uses the Model Context Protocol (MCP), a common set of rules that allows the application to access information and carry out approved actions.
An MCP server can provide:
- Tools for actions such as searching a database or creating an event;
- Resources that provide context, such as files or records; and
- Prompts that provide reusable templates for specific tasks.
For example, the CookieYes MCP Server allows compatible AI assistants to check banner status, retrieve embed codes, trigger cookie scans, and manage banner settings.
MCP servers may run locally or remotely. Access controls, user approval, and secure authentication are important when a server handles personal information or acts on a user's behalf.
