Our in-product RAG chatbot has two halves. The first is retrieval — hybrid BM25 + bge-m3 + a fine-tuned reranker, running locally as native . NET/ONNX, with zero API surface to anyone.
Source: [Dev.to](https://dev.to/wuicframework/running-the-wuic-assistant-on-a-local-llm-ollama-an-mcp-server-and-a-free-agentic-vs-code-4233)