Semantic Kernel
A model-agnostic SDK for building, orchestrating and deploying AI agents and multi-agent systems, with support for Python, .NET and Java.
Semantic Kernel is a model-agnostic SDK that lets developers build, orchestrate and deploy AI agents and multi-agent systems, from a single chatbot to a complex multi-agent workflow. It connects to a range of model providers, including OpenAI, Azure OpenAI, Hugging Face and NVIDIA, and can run models locally through Ollama, LM Studio or ONNX.
The SDK provides an agent framework for building modular agents with access to tools, plugins, memory and planning, an orchestration layer for coordinating specialist agents in multi-agent workflows, and a process framework for modeling complex business processes as structured workflows. Its plugin ecosystem extends agents with native code functions, prompt templates, OpenAPI specifications or the Model Context Protocol, and it integrates with vector databases including Azure AI Search, Elasticsearch and Chroma, plus multimodal input for text, vision and audio.
Semantic Kernel requires Python 3.10 or later, .NET 10 or later, or JDK 17 or later, and runs on Windows, macOS and Linux. Microsoft has since positioned Microsoft Agent Framework as its enterprise-focused successor.