Three self-contained modules. Each one is a real, runnable project with a companion article.
Build a search that understands meaning, not just keywords. Uses Microsoft.Extensions.AI, Ollama for local embeddings, and pgvector on Neon Serverless Postgres to store and query vectors.
Ground your AI responses in real data. Ingest documents, embed them, retrieve the relevant chunks at query time, and feed them to an LLM - so the AI answers from your data, not from guesswork.
Expose your .NET services as tools that any AI client can call. GitHub Copilot, Claude, and Cursor connect to your server and use your data - without you rewriting anything per client.
Enter your email and I'll send you the full source code for all three modules - Semantic Search, RAG System, and MCP Server - ready to clone and run.