Anthropic unveils new AI framework
Anthropic has announced the open-sourcing of the Model Context Protocol (MCP), a new standard designed to bridge the gap between AI systems and the data they rely on. By enabling secure, two-way connections between AI tools and external data sources such as business tools, content repositories, and development environments, MCP aims to address challenges posed by fragmented integrations and information silos.
The Model Context Protocol
MCP introduces a universal framework for integrating AI assistants with diverse datasets. Its core components include:
- MCP Specification and SDKs: These provide developers with the tools to create both MCP servers (to expose data) and MCP clients (AI applications that connect to these servers).
- Local MCP Server Support: Available through Claude Desktop apps, this feature allows developers to test connections locally.
- Open-source Repository: Pre-built MCP servers for platforms like Google Drive, Slack, GitHub, and Puppeteer are available for immediate use.
The protocol is designed to replace custom connectors for each data source with a standardized approach. As the ecosystem evolves, MCP is expected to enable AI systems to maintain context across various tools and datasets, creating a more cohesive architecture.
Availability
Developers can start using MCP immediately by installing pre-built servers via the Claude Desktop app or contributing to the open-source repository. Existing Claude for Work customers can test local MCP servers now, with remote production server support expected soon. Subscribers to Anthropic’s Claude Enterprise plan can leverage MCP to connect Claude chatbots directly with internal systems.
About Anthropic
Anthropic is a leading AI safety and research company known for its Claude family of language models. The introduction of MCP aligns with its mission to develop safer and more capable AI systems while fostering collaboration through open-source initiatives.