microsoft/semantic-kernel

What is it?

Key features

• 01Connects to OpenAI, Azure OpenAI, Hugging Face, Ollama, LM Studio and ONNX models.
• 02Builds agents that use tools, memory and planning.
• 03Orchestrates multiple specialist agents within one workflow.
• 04Extends agents with native functions, prompt templates, OpenAPI specs and MCP.
• 05Integrates with Azure AI Search, Elasticsearch and Chroma as vector stores.
• 06Supports text, vision and audio inputs.

Use cases

• Adding a chatbot or assistant that calls internal APIs to an existing .NET or Python application.
• Building a triage system where several specialist agents handle billing, refunds or support requests.
• Running a model locally with Ollama or ONNX while keeping the same application code.
• Combining LLM calls with retrieval from a vector database for question answering over documents.

Quick start

1. 1

   Set your OpenAI or Azure OpenAI API key as an environment variable

   export OPENAI_API_KEY=sk-...
2. 2

   Install the Python package

   pip install semantic-kernel
3. 3

   Run the basic Python agent example from the README

Things to know

• !Semantic Kernel has been succeeded by Microsoft Agent Framework, so the README points new users there.

Topics

FAQ

What is semantic-kernel?

An SDK for building AI agents and multi-agent systems in Python, .NET and Java, with model, plugin and vector database integrations. Semantic Kernel is a model-agnostic SDK for building, orchestrating and deploying AI agents and multi-agent systems. It targets Python, .NET and Java, and runs on Windows, macOS and Linux.

Is semantic-kernel free to use? What license does it have?

semantic-kernel is open source under the MIT license. Check the license terms before using it in commercial products.

Is semantic-kernel still maintained?

Actively maintained. The last commit to semantic-kernel was on Oct 20, 2026.

What can I use semantic-kernel for?

Adding a chatbot or assistant that calls internal APIs to an existing .NET or Python application. Building a triage system where several specialist agents handle billing, refunds or support requests. Running a model locally with Ollama or ONNX while keeping the same application code. Combining LLM calls with retrieval from a vector database for question answering over documents.

How do I install semantic-kernel?

Set your OpenAI or Azure OpenAI API key as an environment variable: export OPENAI_API_KEY=sk-... Install the Python package: pip install semantic-kernel Run the basic Python agent example from the README

Where can I get semantic-kernel?

The source code and installation instructions are on GitHub: https://github.com/microsoft/semantic-kernel