LangGraph Blog Agent 📝

An AI-powered blog generation system using LangGraph, Langchain, and Groq to research and write high-quality technical blog posts. The agent utilizes Tavily for real-time web research and operates in an interactive Streamlit application.

🌟 Features

• Multi-agent LangGraph Workflow: Dynamically routes between research and orchestration modes based on topic requirements.
• Automated Web Research: Leverages Tavily Search to gather up-to-date evidence and context.
• Dynamic Planning: An Orchestrator node creates a structured plan with specific sections, goals, and target word counts.
• Parallel Processing: Fan-out architecture delegates section writing to parallel worker nodes for faster generation.
• Interactive UI: A Streamlit interface to input topics, watch the agent's real-time execution log (including queries, evidence, and plans), and view/download generated Markdown blogs.

🏗️ Architecture Workflow

The system operates on a state graph defined by the following Mermaid diagram:

graph TD
    User([User Input: Topic]) --> Router[Router Node: Determine Mode]
    Router -->|needs_research = true| Research[Research Node: Tavily Search]
    Router -->|needs_research = false| Orchestrator[Orchestrator Node: Planning]
    
    Research --> Orchestrator
    
    Orchestrator -->|fanout tasks| Worker[Worker Node: Section Writing]
    
    Worker --> Reducer[Reducer Node: Assembly]
    Reducer --> Output([Generated Markdown Blog])

Loading

🚀 Prerequisites

• Python >= 3.12
• Groq API Key for LLM inference (using models like llama-3.3-70b-versatile).
• Tavily API Key for automated web research.

📦 Installation

This project uses uv for fast dependency management, but standard pip can also be used.

1. Clone the repository:

   git clone https://github.com/yourusername/langgraph-blog-agent.git
   cd langgraph-blog-agent

2. Environment Variables: Create a .env file in the root directory and add your API keys:

   GROQ_API_KEY=your_groq_api_key_here
   TAVILY_API_KEY=your_tavily_api_key_here

3. Install dependencies: If using uv:

   uv pip install -r pyproject.toml

   Or using standard pip:

   pip install dotenv ipykernel langchain-groq langchain-tavily langgraph streamlit

🎮 Usage

Start the Streamlit application:

streamlit run app.py

1. Open your browser to the URL provided by Streamlit (usually http://localhost:8501).
2. Navigate to the 🚀 Generate New Blog tab.
3. Enter a topic (e.g., "The Future of Multi-Agent Systems") and click "Generate Blog".
4. Watch the agent's execution log as it routes, researches, plans, and writes.
5. Go to the 📚 View & Manage Blogs tab to read or download the final Markdown file.

📁 Project Structure

• app.py: The main Streamlit web interface and agent execution visualizer.
• src/: Core agent logic.
  • graph.py: LangGraph state graph definition mapping nodes and edges.
  • nodes.py: Implementation of individual agent nodes (Router, Research, Orchestrator, Worker, Reducer).
  • schemas.py: Pydantic models for structured LLM outputs and Agent State.
  • prompts.py: System prompts for guiding the LLM at each node stage.
• generated_blogs/: Output directory where the final Markdown blog posts are saved.