🤖 Blog Generator Using Agentic AI

• An AI-powered multi-agent system that researches, plans, and writes complete blog posts automatically.

• Built using LangGraph, LangChain, and Streamlit, this project demonstrates how LLM agents can collaborate in a structured workflow to generate high-quality blog content.

✨ Features

• Agent-based architecture
• Automated research
• Structured blog planning
• Parallel section generation
• Evidence collection with citations
• Permanent blog memory
• Live Markdown preview
• Download blog as Markdown
• Execution logs & workflow tracking

🏛 Architecture

The system is built as a LangGraph workflow:

🛠 Tech Stack

• LLM Orchestration: LangGraph
• LLM Framework: LangChain
• UI: Streamlit
• Research Tool: Tavily
• Data Handling: Pandas

⚙️ How It Works

• User enters a blog topic.
• Router agent decides if research is needed.
• Research agent gathers evidence.
• Orchestrator creates a structured blog plan.
• Worker agents generate blog sections.
• Reducer merges sections into a final Markdown blog.
• Blog is saved locally for future access.

🚀 Running Locally

1. Clone the repository

git clone https://github.com/saparya05/Blog-Generator-Using-Agentic-AI
cd AI-blog-writing-agent

2. Clone the repository

pip install -r requirements.txt

3. Add environment variables

Create a .env file in the project root with the following:

OPENAI_API_KEY=your_openai_key
TAVILY_API_KEY=your_tavily_key

4. Run the app

streamlit run frontend.py

📜 License

This project is open-source and available under the MIT License.