Overview
Semantic Kernel is a model-agnostic SDK designed to empower developers in building, orchestrating, and deploying AI agents and multi-agent systems. It supports integration with various LLMs, including OpenAI, Azure OpenAI, Hugging Face, and Nvidia. The framework allows developers to create modular AI agents with capabilities such as tool/plugin access, memory management, and planning. It facilitates the orchestration of complex workflows using collaborating specialist agents. Semantic Kernel supports extensions through native code functions, prompt templates, OpenAPI specs, and the Model Context Protocol (MCP). It offers seamless integration with vector databases like Azure AI Search and Elasticsearch. The framework is built for enterprise-grade reliability, observability, and security with stable APIs. It supports local deployment with Ollama, LMStudio, or ONNX and can model complex business processes with a structured workflow approach.