-
Feature
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
False
-
-
False
-
Not Selected
Feature Overview
The Model Context Protocol (MCP), is an open protocol that enables seamless integration between LLM applications and external data sources and tools. This protocol connects AI-powered assistants to the systems where data lives, including content repositories, business tools, and development environments.
This card is for exploring the adoption or possible alignment with MCP or the concepts of MCP as a mechanism to extend RHEL AI RAG artifacts and future RHEL AI Agentic AI artifacts or integrations.
Goals
- Research viability and Model Context Protocol (MCP) as a protocol for integration of RHEL AI RAG or Agentic artifacts with external data sources
- Recommendations for moving forward or not in the adoption MCP or other similar industry protocols to integrate LLMs with external applications or systems.
Requirements
- Report with recommended next steps in the adoption or for passing in the adoption of MCP
- Recommendation of other potential protocols to consider for the adoption in RHEL AI for RAG or Agentic artifacts.
Done
- Delivery of report with recommendations
Use Cases
Applications using LLMs for chat or virtual assistant interfaces or AI workflows require interconnections with sources (tools, applications, databases, and other agents) with the correct context they need to complete their tasks.
Antrophic's Model Context Protocol (MCP) is one of the most recent attempts to make this possible. In this case, the protocol is already being used by well-known organizations delivering LLM-powered services and applications.
Announcement: https://www.anthropic.com/news/model-context-protocol
Out of Scope
- The implementation of a productized MCP support for RHEL AI (that will be considered in a future card if the reports recommends the adoption of MCP)
Documentation Considerations
- No user-facing documents.
- Internal-only research and report
Questions to Answer
- Are there overlapping with LlamaStack API or can these be considered complementary APIs?
*Customer Considerations *
- Is the protocol simple and extensible enough to gain traction in the market?
- How fast do libraries like LangChain, LangGraph, LlamaIndex, Haystack, CrewAI, AutoGen bring support for the MCP? This can be used as an indication of market opportunity and impact.
- …