daveebbelaar/langchain-experiments

Demonstrates RAG patterns and agentic workflows through practical examples including YouTube transcript indexing with FAISS vector search, multi-step chains for sequential LLM calls, and agent loops that make decisions and take actions. Integrates OpenAI models (GPT-3.5 Turbo, GPT-4) with external data sources and APIs (SerpAPI) to extend LLM capabilities beyond simple API calls, using LangChain's modular architecture for memory persistence, prompt optimization, and tool integration.

1,130 stars. No commits in the last 6 months.