MCP (Model Context Protocol)

The Model Context Protocol (MCP) is an open standard initiated by Anthropic that standardises the connection of AI models to data in local or cloud-based systems. Access is secure when implemented with appropriate controls (authorisation, transport, deployment).

---

Core concept

MCP acts as a universal connector between LLMs and tools (such as databases, Git repositories, or APIs). It replaces proprietary integrations with an open ecosystem of servers and clients.

Relevance

• AI Agent Architecture: Use of MCP to control development tools through AI.
• Data Sovereignty: Control over which data contexts are made available to a model via MCP servers.
• Standardised Tool Integration: Rapid connection of new data sources to AI assistants without custom code.

---

Related topics

• Technology: AI Development, the development context for AI-assisted work with MCP (Model Context Protocol).
• Innovation: GenAI, the pipeline in which MCP (Model Context Protocol) is applied.
• Standards, the standards section that frames MCP (Model Context Protocol).