Semantic Kernel
Microsoft's enterprise agent SDK — now transitioning to Agent Framework.
About
Semantic Kernel is Microsoft's model-agnostic SDK for building AI agents and multi-agent systems in Python, .NET, and Java: connect to any LLM provider, wrap native code or APIs as plugins, orchestrate collaborating specialist agents, and hook into vector DBs and MCP servers. Important transition note (verified in the README, Oct 2026): Microsoft is folding Semantic Kernel into Microsoft Agent Framework (v1.0, production-ready, with a migration guide) as its enterprise-ready successor. New projects should evaluate the Agent Framework repo before committing to SK APIs.
What it can do
- **Multi-language SDK**: Python, .NET, and Java with the same concepts
- **Plugins**: expose native code, prompt templates, OpenAPI specs, or MCP servers as tools
- **Multi-agent orchestration**: collaborating specialist agents and process workflows
- **Model flexibility**: OpenAI, Azure OpenAI, Hugging Face, NVIDIA NIM, local (Ollama/LMStudio)
- **Enterprise posture**: observability, security, and stable-API commitments
- **Migration path**: official guide to Microsoft Agent Framework v1.0
Setup
```bash
pip install semantic-kernel
```
```python
from semantic_kernel import Kernel
kernel = Kernel() # add chat service + plugins, then invoke
```
.NET: dotnet add package Microsoft.SemanticKernel. See the migration guide at learn.microsoft.com (agent-framework) before starting new projects.
Prerequisites
Python 3.10+ (or .NET 10+ / JDK 17+). An LLM provider API key (OpenAI, Azure OpenAI, Hugging Face, etc.).
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: 28,655 GitHub stars (log-scaled vs catalog peers); last push 0d ago; 39 releases/12mo; 353 open issues. Manual review: Microsoft's enterprise agent SDK (Py/.NET/Java), 28.6k stars — BUT README banner: transitioning to Microsoft Agent Framework v1.0 successor; evaluate the new repo + migration guide before starting new projects. License: MIT.
★ 28,655 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 8.3/10
- Maintenance · recency, releases, issue load
- 9.5/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · 28,655 GitHub stars (log-scaled vs catalog peers)
- · last push 0d ago; 39 releases/12mo; 353 open issues
- · manual review: Microsoft's enterprise agent SDK (Py/.NET/Java), 28.6k stars — BUT README banner: transitioning to Microsoft Agent Framework v1.0 successor; evaluate the new repo + migration guide before starting new projects
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.