The watsonx MCP Server acts as a crucial bridge, connecting Claude Code (Opus 4.5) with the powerful AI capabilities of IBM watsonx.ai. It allows Claude to delegate specific AI workloads to a range of IBM's foundation models, including Granite, Llama, and Mistral, for tasks such as text generation, chat, and embeddings. This creates a versatile two-agent system where Claude handles high-level reasoning, while watsonx.ai provides specialized functionalities like batch inference on enterprise data, advanced document analysis, and sophisticated RAG (Retrieval Augmented Generation) pipelines.
Key Features
01Create text embeddings for semantic understanding
020 GitHub stars
03List all available watsonx.ai foundation models
04Analyze documents with summarization, Q&A, and semantic search
05Generate text using watsonx.ai foundation models
06Engage in conversations with watsonx.ai chat models
Use Cases
01Analyze and query external drive documents for summarization, entity extraction, and Q&A.
02Delegate specialized AI tasks from Claude to IBM-specific models for enhanced capabilities.
03Perform batch inference on enterprise data and generate embeddings for RAG pipelines.
