🧠Semantic Kernel
A comprehensive collection of guides and tutorials on Microsoft Semantic Kernel - the open-source SDK that enables you to build AI agents and integrate large language models into your applications with ease.
Semantic Kernel - An Overview
What Semantic Kernel is, how it works, which languages it supports, and where I'm using it.
About Semantic Kernel
Semantic Kernel is an open-source SDK that lets you easily build agents that can call your existing code. As a highly extensible SDK, you can use Semantic Kernel with models from OpenAI, Azure OpenAI, Hugging Face, and more! By combining your existing C#, Python, and Java code with these models, you can build agents that answer questions and automate processes.
Key Semantic Kernel Concepts
đź’¬ LLM Chat Completion
Build conversational AI experiences using large language models with streaming and context management.
🔍 Semantic Index
Leverage semantic search and retrieval capabilities to find relevant information using vector embeddings.
- Semantic Index What Is It?
- Vector Search Basics — Cosine Similarity and ANN Indexes (HNSW, IVF, PQ)
- Retrieval‑augmented generation (RAG) patterns and evaluation
- Chunking strategies, metadata filters, and hybrid search (BM25 + vectors)
- Semantic Index with Azure AI Search
- Semantic Index with Qdrant
- Semantic Index with Azure Cosmos DB
- Semantic Index with SQL Server & Azure SQL
- Copilot Retrieval API (Beta): What You Can Do Today
⚡ Kernel Functions
Create reusable AI capabilities by combining prompts and native code as kernel functions.
🧩 Memory
Store and retrieve contextual information to build stateful AI applications with long-term memory.
🎯 Triggers
Automate AI workflows with event-driven triggers and scheduled executions.
🔄 Orchestrations
Chain multiple AI operations together to create complex, multi-step intelligent workflows.
- Semantic Kernel Orchestration
- Outlook Agent — Drafts, Meetings & Calendar Blocks
- Knowledge Agent — unified access to mail, OneDrive and SharePoint
- Agent-to-Agent protocol (A2A) — orchestrating across agents
- Validation Agent — prove results and verify correctness
- Human-in-the-Loop — design patterns and implementation guidance
