Self-hosted ML inference MCP server providing embeddings and summarization tools using sentence-transformers and Ollama, with low-latency preloaded models and API-key authentication.
