LangGraph Agentic Systems — Learning & Reference Hub

---

Purpose

This repository serves as a small, focused learning space for developers exploring how to build agentic systems using LangGraph. It provides working examples for patterns such as tool integration, RAG pipelines, persistent memory, Streamlit frontends, and MCP connectivity.

The intent is educational: the code functions as a set of reference templates, not a packaged production system.

---

Learning Outcomes

✔ Transition from linear prompting to graph-based execution ✔ Build RAG pipelines using FAISS for grounded responses ✔ Attach real-world tools (search, calculator, stock prices) ✔ Wrap agent workflows in Streamlit-driven demos ✔ Connect external processes via MCP ✔ Use STM/LTM memory for multi-session continuity ✔ Trace workflows with LangSmith for inspection and debugging

---

Module Breakdown

📌 1. Graph Orchestration

langgraph_backend.py Branching logic, loops, corrective flow, state tracking.

---

📌 2. Retrieval-Augmented Generation (RAG)

langgraph_rag_backend.py PDF ingestion → embeddings → FAISS similarity → grounded answers.

---

📌 3. Tool Execution

langgraph_tool_backend.py Search, calculator, stock price querying via agent tool calls.

---

📌 4. Model Context Protocol (MCP)

langgraph_mcp_backend.py Connecting to external processes through a standard protocol.

---

📌 5. Memory & Persistence

langgraph_database_backend.py SQLite checkpointing for resume + HITL workflows. Notebooks demonstrate STM vs LTM strategies and summarization.

---

Repository Structure

📂 Project
│
├── *.ipynb                       # Notebook tutorials (STM, LTM, HITL)
├── langgraph_*.py                # Backends (Agents, RAG, Tools, MCP, Persistence)
├── streamlit_*.py                # Frontend UI demos
├── mcp_server.py                 # MCP server entry point
├── chatbot.db                    # SQLite persistence layer
├── requirements.txt              # Dependencies
├── .env                          # Environment variables (ignored)
├── .gitignore
└── LICENSE

---

Installation & Environment

Requirements: • Python ≥ 3.10 • Git

Setup:

git clone <repository-url>
cd chatbot
python -m venv venv

Activate:

Windows:

venv\Scripts\activate

macOS/Linux:

source venv/bin/activate

Install:

pip install -r requirements.txt

---

Environment Configuration

.env file:

OPENAI_API_KEY=...
LANGSMITH_TRACING=true
LANGSMITH_API_KEY=...
LANGSMITH_PROJECT=chatbot-project

Optional:

GROQ_API_KEY=...
TAVILY_API_KEY=...

---

Running Modules

RAG demo:

streamlit run streamlit_rag_frontend.py

Tools demo:

streamlit run streamlit_frontend_tool.py

Streaming:

streamlit run streamlit_frontend_streaming.py

Basic:

streamlit run streamlit_frontend.py

---

Tech Stack

🧠 Reasoning & Agent Orchestration

📚 Retrieval & Embeddings

🧰 Tools & Connectivity

Tools demonstrated: Search, Calculator, Stock APIs

🗄 Persistence & State

🖥 Frontend Demos

📊 Tracing & Evaluation

---

License

Licensed under the MIT License.