// the find
NirDiamant/GenAI_Agents
50+ tutorials and implementations for Generative AI Agent techniques, from basic conversational bots to complex multi-agent systems.
A collection of 52+ Jupyter notebooks walking through generative AI agent patterns, from a simple LangChain chatbot up to multi-agent systems with LangGraph, AutoGen, and CrewAI. It's an educational resource, not a library or framework — nothing here is importable. The target audience is developers learning to build agents for the first time.
Coverage is genuinely wide: it hits LangGraph state machines, MCP integration, AutoGen multi-agent collaboration, and PydanticAI alternatives side by side, so you can compare patterns without hunting across different repos. The LangGraph tutorials in particular show the state graph mechanics clearly, with each node and edge made explicit rather than hidden behind magic. Several notebooks come with YouTube walkthroughs and blog posts, which is useful when a notebook alone doesn't explain the design intent. Recent activity (June 2026) means the MCP and newer framework tutorials aren't stale.
Everything runs against OpenAI by default — API keys, model names, and cost assumptions are baked into almost every notebook, so adapting to another provider means editing each one manually with no shared config. Quality across 52 notebooks is uneven; the beginner entries are thin wrappers that add little over the official LangChain docs, while the advanced ones (ClauseAI, ATLAS) require external services like Pinecone that you have to set up yourself. No shared utilities, no tests, no CI — if a dependency breaks silently, you won't know until the notebook fails mid-run. The repo is also drifting toward being a funnel for the author's paid course and newsletter, which is fine, but it means the README is increasingly marketing rather than navigation.