Building an AI Chatbot with LangGraph

I’ve been developing an AI chatbot using Python, LangGraph, LangChain, OpenAI, MCP, RAG, SQLite, and Streamlit, focusing on building a stateful and tool-enabled AI application.

🔹 Built a LangGraph workflow with nodes, edges, and state management
🔹 Implemented TypedDict and add_messages for structured conversation state
🔹 Integrated OpenAI models through LangChain
🔹 Implemented conversation persistence using LangGraph checkpointers
🔹 Added SQLite-based persistent memory for maintaining conversations
🔹 Implemented unique thread_id management for multiple conversations
🔹 Added conversation history and thread switching through Streamlit
🔹 Implemented streaming AI responses for an interactive chat experience
🔹 Integrated tool calling to allow the chatbot to interact with external capabilities
🔹 Implemented MCP (Model Context Protocol) for connecting the AI with external tools and resources
🔹 Implemented RAG (Retrieval-Augmented Generation) for retrieving relevant information from knowledge sources
🔹 Integrated vector-based semantic search for context retrieval
🔹 Built agentic workflows that combine LLMs, state, tools, memory, and retrieved context

Through this project, I’m gaining hands-on experience in LLM application development, agentic AI, graph-based orchestration, tool calling, MCP, RAG, vector search, memory, state management, and persistent AI workflows.

Tech Stack:
Python | LangGraph | LangChain | OpenAI | MCP | RAG | Vector Database | SQLite | Streamlit

Development is ongoing, with continuous improvements and new capabilities being added as I explore production-oriented approaches to building intelligent AI applications.