license	title	sdk	emoji	colorFrom	colorTo	short_description
apache-2.0	Chatbot	streamlit	🚀	red	yellow	LangGraph Chatbot

✨Agentic AI Chatbot with LangGraph🧑‍💻

📌 Project Overview

This project is a Streamlit-based Agentic AI application that demonstrates graph-driven AI workflows using LangGraph and LangChain, powered by GROQ LLMs and integrated external tools.

It showcases how modern agentic systems can reason, route, and act across multiple capabilities instead of behaving like a single static chatbot.

The application includes three core use cases:

• 💬 Basic Chatbot
  A simple conversational agent powered by an LLM.

• 🌐 Chatbot with Web Tools
  A tool-augmented agent capable of invoking web/search tools when required.

• 📰 AI News Agent
  Fetches recent AI-related news using Tavily, summarizes it with an LLM, and saves structured Markdown reports.

All core logic lives under src/langgraphagenticai/, while app.py bootstraps the Streamlit UI.

---

🧰 Tech Stack

• 🐍 Python 3.13 (virtual environment in chatbot_env/)
• 🎛 Streamlit – interactive UI for chat and results
• 🔗 LangGraph + LangChain – graph-based agent orchestration
• ⚡ GROQ LLM – fast LLM inference backend
  (configured in src/langgraphagenticai/LLMS/groqllm.py)
• 🔍 Tavily API – AI news search

🔐 Environment variables (GROQ, Tavily API keys) are loaded from .env.
Do not commit secrets.

---

🏗 Architecture

🔁 High-Level Flow

• app.py

  • Launches Streamlit
  • Loads UI controls
  • Reads user input and use-case selection
  • Configures LLMs
  • Builds and executes the LangGraph workflow

• graph_builder.py

  • Defines graph structures per use case
  • Compiles graphs into executable workflows

• nodes/

  • Contains task-specific agent nodes
  • Each node exposes callable methods used as graph steps

• ui/streamlitui/

  • UI helpers for inputs and result rendering

---

🧠 Graph Composition (Per Use Case)

💬 Basic Chatbot

• Single node: BasicChatbotNode.process
• Flow: START → Chatbot → END

🌐 Chatbot with Web

• Chatbot node + Tool node
• Conditional edges route between reasoning and tool execution

📰 AI News Pipeline

A structured three-step agentic workflow:

1. AINewsNode.fetch_news → Tavily search
2. AINewsNode.summarize_news → LLM summarization
3. AINewsNode.save_result → Save Markdown to
   AINews/<frequency>_summary.md

---

🖥 Data & UI Handling

• 🎚 User input is collected via Streamlit controls (LoadStreamlitUI)
• 📄 Results are rendered using DisplayResultStreamlit
  • Chat messages for chatbot modes
  • Markdown previews for AI News summaries
• 🗂 AI News outputs are stored in the AINews/ directory

---

📂 Directory Highlights

• app.py
• src/langgraphagenticai/graph/graph_builder.py
• src/langgraphagenticai/nodes/ai_news_node.py
• src/langgraphagenticai/ui/streamlitui/display_result.py

---

▶️ Running Locally

1️⃣ Create and activate a virtual environment (optional if using chatbot_env/)

2️⃣ Install dependencies:

pip install -r requirements.txt

3️⃣ Set required environment variables in .env (example: GROQ and Tavily API keys)

4️⃣ Start the application:

streamlit run app.py

🪟 Windows Encoding Fix (if needed)

If you encounter encoding issues on Windows:

$env:PYTHONUTF8=1
streamlit run app.py