OSS TanbouOn this page
Overview
Semantic Kernel is a model-agnostic SDK for building agents from AI models, application functions, memory, and orchestrated processes. It connects to providers such as OpenAI, Azure OpenAI, and Hugging Face and supports designs ranging from a single assistant to collaborating specialist agents. The project targets .NET, Python, and Java and can extend agents through MCP and OpenAPI.
Features and best fit
Based on official documentation; not hands-on tested · Content checked:
Key features
Compose models, plugins, and memory into agent behavior
Built-in connectors let applications choose among LLM providers and expose native functions, prompts, OpenAPI operations, or MCP capabilities as agent tools. Vector database integrations, text, vision, and audio handling, multi-agent collaboration, and structured business processes are available as parts of the same SDK.
Sources: [1]
Best fit
Fits teams adding agent behavior to existing applications incrementally
It fits teams that want to start with chat responses and then add business-function plugins, retrieval from vector stores, and handoffs between specialist agents. It is also relevant when an application should control orchestration in .NET or Python code without binding its whole design to one model provider.
Sources: [1]
Before adoption
Supported runtimes and external AI service configuration are required
The README lists Python 3.10+, .NET 10.0+, and Java 17+ as runtime requirements. Even the basic examples require Azure OpenAI or OpenAI credentials and an available model, so credentials must be managed outside source code. Ollama, LMStudio, and ONNX provide local options, but available features and model quality vary by configuration. The project uses the MIT License.
New adopters should compare the successor Microsoft Agent Framework
The pinned README identifies Microsoft Agent Framework as the production-ready successor to Semantic Kernel and links to a migration guide. Teams should distinguish maintenance of existing Semantic Kernel assets from selection of a new agent foundation, then verify API stability, long-term support, and migration timing in official guidance. The latest checked stable .NET release is dotnet-1.80.1.
Official sources
- [1]microsoft/semantic-kernel dotnet-1.80.1 — README(2026-10-04)
- [2]Semantic Kernel dotnet-1.80.1 release(2026-10-04)
- [3]microsoft/semantic-kernel dotnet-1.80.1 — LICENSE(2026-10-04)
Supplemental curator note
Semantic Kernel brings model calls, tools, memory, and workflow orchestration into one application SDK. For a new adoption, compare it with the successor Microsoft Agent Framework before committing to a maintenance and migration path.
Try it in 3 steps
- 1
Create an isolated Python environment
Use Python 3.10 or newer and isolate dependencies from the existing environment.
python3 -m venv sk-demo && . sk-demo/bin/activate - 2
Install Semantic Kernel
Install the official Python package in the isolated environment.
python -m pip install semantic-kernel - 3
Create a Kernel without credentials
Verify the Kernel import and empty service registry without connecting to a model or paid API.
python -c "from semantic_kernel import Kernel; k = Kernel(); print(type(k).__name__, len(k.services))"
Growth
Growth trends · Last 30 days
28,624 Stars
Trend data is still being collected.
Development activity
Last 90 days · weekly
- Commits (last 30 days)
- 13
- Open PRs
- 217
Development activity is still being collected.
Built with
Categories and tags
GitHub data
GitHub dataView detailed GitHub data
GitHub Topics
- ai
- artificial-intelligence
- llm
- openai
- sdk
Related information
Write a related articleShare a guide or use case for this OSS in Markdown. Articles are published after administrator approval.
Explore next
- Frontman709 Stars
2 shared tag(s) · 2 shared category(s)
Trace selected browser UI through live DOM, CSS, and source maps and let an AI agent edit the owning source
ReScript - Dify157,795 Stars
2 shared tag(s) · 1 shared category(s)
design and publish AI workflows, retrieval, and agents on one platform
TypeScript - Langflow155,490 Stars
2 shared tag(s) · 1 shared category(s)
make AI workflows visible, then publish them as APIs and MCP tools
Python - Flowise55,487 Stars
2 shared tag(s) · 1 shared category(s)
treat the visual AI workflow platform as an EOL migration and maintenance target
TypeScript - LlamaIndex52,404 Stars
2 shared tag(s) · 1 shared category(s)
compose ingestion, retrieval, RAG, and agents from modular packages
Python - Cherry Studio52,354 Stars
2 shared tag(s) · 1 shared category(s)
a desktop AI workspace for multi-provider LLMs, agents, knowledge/RAG, MCP, and a local API gateway
TypeScript
Report incorrect information
Tell us if any listing information is incorrect or outdated.