Autonomous Blog Team (CrewAI Project)

An autonomous agent team built with the crewai library that researches any given technical topic on the web and creates a professional blog post based on that research.

Team & Process

The project uses two agents working in a sequential process (Process.sequential):

1. Technical Research Specialist (Researcher)

   • Tools: SerperDevTool (Web Search), ScrapeWebsiteTool (Website Scraper), verify_source (Custom Source Validator)
   • Task: Finds 3-5 reliable sources, validates them, reads their content, and prepares a technical research report

2. Technical Content Writer (Writer)

   • Task: Takes the researcher's report as context and transforms it into an 800-1000 word, SEO-friendly, fluent blog post

Prerequisites

• Python 3.8+
• OpenAI API Key
• Serper API Key (for web search via serper.dev)

Installation

# Clone the repository
git clone https://github.com/AbdulSametTurkmenoglu/autonomous_blog_team_crew_ai.git
cd autonomous_blog_team_crew_ai


# Create virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate  # Windows: .\.venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Setup environment variables
cp .env.example .env  # Windows: copy .env.example .env
# Edit .env and add your API keys

.env file contents:

OPENAI_API_KEY="sk-..."
SERPER_API_KEY="..."

Note: Get your SERPER_API_KEY from serper.dev

Usage

The project runs via the run.py script.

Run with default topic:

python run.py

The script will process the default topic: Direct Preference Optimization (DPO)

Run with custom topic:

python run.py "What is LangGraph and how does it work?"

When the team starts working, you'll see detailed logs in the terminal (enabled by verbose=True). Once completed, a markdown file named blog_ciktisi_...md will be created.

How It Works

1. Research Phase:

   • Researcher agent searches the web for relevant sources
   • Validates source credibility
   • Scrapes and analyzes content
   • Compiles a comprehensive research report

2. Writing Phase:

   • Writer agent receives the research report
   • Transforms technical information into engaging content
   • Ensures SEO optimization
   • Produces a polished 800-1000 word blog post

Features

• Autonomous Research: Agents independently find and verify sources
• Source Validation: Custom tool ensures information credibility
• SEO Optimization: Generated content is search-engine friendly
• Sequential Process: Clear workflow from research to final output
• Detailed Logging: Verbose mode shows the entire decision-making process
• Markdown Output: Ready-to-publish blog posts in markdown format