Text to SQL using LLMs #4 - Removing the Router Agent & Adding Modular Nodes (v5) | LangGraph

In this video, I walk through version 5 of my Text-to-SQL app, where I remove the Table Router Agent and introduce modular LLM nodes for subquestion generation, table mapping, and column selection. All prompts are now standardized with structured JSON I/O for better consistency and reasoning. (Note: Version 4.1 updates - richer table and column context in the Knowledge Base Builder, are not covered in this video but are part of the project’s backend improvements.) 🔗 GitHub Repo: https://github.com/v1jaysundaram/text2sql 🎥 Full Playlist: https://youtube.com/playlist?list=PLEHeb1HGikpQJvPlZEyrMLOjopsV5Ui_3&si=8dqEScL-uSyIskqx 🤝 Connect with me on LinkedIn: https://www.linkedin.com/in/vijay-sundaram/