Semantic Kernel

Semantic Kernel is an SDK for developers building AI agents and coordinated workflows inside applications. It lets developers choose local model backends or cloud providers for those applications.

• Agents and workflows: Build agents with tools, memory and planning, coordinate specialist agents, or define structured business workflows with the Process Framework.
• Plugins and services: Connect agents to native code functions, prompt templates, OpenAPI services and Model Context Protocol (MCP) tools.
• Stored information: Integrate vector databases and search services such as Azure AI Search, Elasticsearch and Chroma.
• Text, vision and audio: Build applications that process these input types through compatible model backends.

The SDK supports C#, Python and Java on Windows, macOS and Linux. Local deployment works with Ollama, LMStudio or ONNX. Connections to OpenAI and Azure OpenAI send model requests to those cloud services, so the choice of backend determines where model processing happens.

Semantic Kernel is open source under the MIT license. Microsoft identifies Microsoft Agent Framework as its successor and provides a migration guide for existing Semantic Kernel applications.