knowledge-support-ai-agent

Knowledge Support AI Agent is described as a production-grade reference implementation of a conversational AI support agent. It is built to demonstrate how to structure a real-world AI backend using Clean Architecture, with RAG, semantic memory, persistent chat history, and provider independence included from the start. The project states that business logic stays isolated from LLM providers, databases, and messaging platforms, making every layer independently testable and replaceable. WhatsApp Cloud API is the intended communication channel, and a REST API is available for direct integration and local development. The system includes features such as conversational chat with persistent history per contact, RAG retrieval via semantic search, hybrid search combining vector similarity and PostgreSQL full-text search, document ingestion with pgvector indexing, tool calling, conversation history optimization, query rewriting, context compression, provider independence, OpenTelemetry instrumentation, rate limiting, error handling, and security headers. It also provides REST API and CLI interfaces.