Semantic Kernel
Semantic Kernel is a model-agnostic Microsoft SDK for integrating AI agents and multi-agent systems into your own applications – in C#/.NET, Python and Java. Microsoft now describes Microsoft Agent Framework as its successor.
- Project
- Semantic Kernel
- Publisher
- Microsoft
- Repository
- github.com/microsoft/semantic-kernel
- Documentation
- learn.microsoft.com/en-us/semantic-kernel/
- Languages
- C#, Python, Java
- Category
- AI agent frameworks
What is Semantic Kernel?
Semantic Kernel is an open-source SDK (MIT licence) that developers use to integrate language models, agents and multi-agent workflows into existing applications. It supports Python, .NET and Java and runs on Windows, macOS and Linux.
As of October 2026, the repository states that Microsoft Agent Framework is the successor to Semantic Kernel and is available at version 1.0 with stable APIs and long-term support. Microsoft provides a migration guide.
Core capabilities
- Model flexibility
- Connections to OpenAI, Azure OpenAI, Hugging Face and NVIDIA, among others, and to local models via Ollama, LM Studio or ONNX.
- Agents and plugins
- Agents with access to tools, memory and planning; plugins from your own code, prompt templates, OpenAPI specifications or MCP.
- Multi-agent systems
- Orchestrating several specialised agents in a shared workflow.
- Process Framework
- Structured modelling of business processes as workflows.
- Vector databases
- Integrations with Azure AI Search, Elasticsearch and Chroma, among others.
When Semantic Kernel fits
Semantic Kernel is an obvious candidate when your applications are built in .NET or Java and existing code is to be exposed to agents as plugins. For new projects, it is worth also evaluating Microsoft Agent Framework because of the stated succession.
Using Semantic Kernel effectively in an AI-first business model
Plugins make existing systems accessible to agents – customer data, quotes or orders, for example. This makes it central to decide which functions an agent may only read, which it may prepare and which it may execute with approval or on its own.
Define these action rights per plugin and document them independently of the SDK. That way accountability and control are preserved, even if you later move to a different foundation such as Microsoft Agent Framework.
From tool to outcome
In AI-first coaching you clarify which outcomes AI agents should be accountable for in your go-to-market and within which framework. REVAGO then designs, builds and runs your system. You pay for results – not for activity.
Frequently asked questions
What is Semantic Kernel?
Semantic Kernel is a model-agnostic open-source SDK by Microsoft (MIT licence) for building and orchestrating AI agents and multi-agent systems in Python, .NET and Java. Existing code can be exposed to agents as plugins.
How do Semantic Kernel and Microsoft Agent Framework relate?
According to the repository, Microsoft Agent Framework is the successor to Semantic Kernel and is available at version 1.0 with stable APIs and long-term support. Microsoft provides a migration guide from Semantic Kernel.
Does Semantic Kernel support MCP?
Yes. Plugins can be added from your own code, prompt templates, OpenAPI specifications and via the Model Context Protocol (MCP), among other sources.
Further reading in the wiki
A neutral overview, not a recommendation of individual providers. REVAGO is not affiliated with the listed projects. Descriptions are based on the projects’ public information; features and licences may change.