# Create project directory
mkdir research-crew && cd research-crew

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install crewai crewai-tools duckduckgo-search beautifulsoup4 requests python-dotenv

# .env
OPENAI_API_KEY=your_openai_api_key_here

# tools.py
from crewai.tools import tool


@tool("Web Search")
def web_search(query: str) -> str:
    """Search the web for current information on a topic.

    Args:
        query: The search query to look up.
    """
    from duckduckgo_search import DDGS

    results = DDGS().text(query, max_results=5)
    formatted = ""
    for r in results:
        formatted += f"**{r['title']}**\n{r['body'][:300]}\nURL: {r['href']}\n\n"
    return formatted if formatted else "No results found."


@tool("Read Webpage")
def read_webpage(url: str) -> str:
    """Read and extract the main content from a webpage.

    Args:
        url: The full URL of the webpage to read.
    """
    import requests
    from bs4 import BeautifulSoup

    try:
        response = requests.get(url, timeout=10, headers={
            "User-Agent": "Mozilla/5.0 (Research Agent)"
        })
        soup = BeautifulSoup(response.content, "html.parser")

        # Remove non-content elements
        for tag in soup(["script", "style", "nav", "footer", "header", "aside"]):
            tag.decompose()

        text = soup.get_text(separator="\n", strip=True)
        return text[:4000]  # Limit to avoid token overflow
    except Exception as e:
        return f"Error reading webpage: {str(e)}"

# agents.py
from crewai import Agent
from tools import web_search, read_webpage


def create_researcher():
    return Agent(
        role="Senior Research Analyst",
        goal="Produce comprehensive, accurate, and well-sourced research briefs on any given topic",
        backstory="""You are a veteran research analyst with 15 years of experience
        at a top-tier consulting firm. You are known for your meticulous approach
        to information gathering -- you never settle for surface-level findings.
        You always cross-reference multiple sources, identify primary data,
        and separate facts from opinions. Your research has informed
        billion-dollar strategic decisions.""",
        tools=[web_search, read_webpage],
        verbose=True,
        allow_delegation=True,
        llm="gpt-4o"
    )


def create_writer():
    return Agent(
        role="Content Strategist & Writer",
        goal="Transform raw research into clear, engaging, and well-structured articles",
        backstory="""You are an award-winning content strategist who has written
        for Wired, MIT Technology Review, and The Verge. You specialize in making
        complex technical topics accessible to a broad audience without dumbing
        them down. Your writing is known for strong narrative structure,
        vivid examples, and clear explanations that respect the reader's
        intelligence.""",
        verbose=True,
        allow_delegation=False,
        llm="gpt-4o"
    )


def create_editor():
    return Agent(
        role="Senior Editor",
        goal="Ensure all content is polished, accurate, logically structured, and publication-ready",
        backstory="""You are a meticulous senior editor with experience at
        The New York Times and Harvard Business Review. You have an eye for
        factual accuracy, logical flow, clarity, and engaging prose. You catch
        inconsistencies others miss, tighten wordy paragraphs, and ensure
        every section earns its place. You make direct edits rather than
        just suggesting changes.""",
        verbose=True,
        allow_delegation=False,
        llm="gpt-4o"
    )

# tasks.py
from crewai import Task


def create_research_task(agent, topic):
    return Task(
        description=f"""Conduct thorough research on the following topic:

        "{topic}"

        Your research must include:
        1. An overview of the current state of this topic
        2. Key players, technologies, or concepts involved
        3. Recent developments or breakthroughs (last 12 months)
        4. Specific data points, statistics, or metrics where available
        5. Different perspectives or debates within the field
        6. Practical implications and real-world applications

        Search for at least 3-5 different queries to get comprehensive coverage.
        Cross-reference claims across multiple sources.""",
        expected_output="""A structured research brief containing:
        - Executive summary (3-4 sentences)
        - Key findings organized by subtopic
        - Specific data points and statistics with source attribution
        - Notable quotes or expert opinions
        - Identified trends and predictions
        Total length: 800-1200 words.""",
        agent=agent
    )


def create_writing_task(agent, topic):
    return Task(
        description=f"""Using the research provided, write a comprehensive article on:

        "{topic}"

        Requirements:
        - Compelling opening that hooks the reader with a surprising fact or scenario
        - Clear structure with descriptive headings
        - Specific examples and data woven naturally into the narrative
        - Accessible language that a technical professional can follow
        - A conclusion that synthesizes key insights and looks ahead
        - 1000-1500 words in length
        - Written in markdown format""",
        expected_output="""A publication-ready article in markdown format with:
        - Engaging title and subtitle
        - Strong opening paragraph
        - 4-6 main sections with clear headings
        - Data points and examples integrated throughout
        - Conclusion with forward-looking perspective
        - Professional, authoritative tone""",
        agent=agent
    )


def create_editing_task(agent):
    return Task(
        description="""Review and edit the article for publication. Perform the following:

        1. **Fact-check**: Verify all claims and data match the original research
        2. **Structure**: Ensure logical flow between sections with smooth transitions
        3. **Clarity**: Simplify any convoluted sentences; ensure each paragraph has a clear purpose
        4. **Engagement**: Strengthen the opening hook and closing call-to-action
        5. **Grammar & Style**: Fix any grammatical errors, improve word choice
        6. **Consistency**: Check for consistent tone, tense, and formatting

        Make direct edits to the article. Do not just list suggestions -- produce the final version.""",
        expected_output="""The final, polished article ready for publication.
        Include a brief editor's note at the top (2-3 sentences) summarizing the key improvements made.""",
        agent=agent
    )

# main.py
import os
from dotenv import load_dotenv
from crewai import Crew, Process
from agents import create_researcher, create_writer, create_editor
from tasks import create_research_task, create_writing_task, create_editing_task

# Load environment variables
load_dotenv()


def run_research_crew(topic: str) -> str:
    """Run the full research crew pipeline on a given topic."""

    # Create agents
    researcher = create_researcher()
    writer = create_writer()
    editor = create_editor()

    # Create tasks
    research_task = create_research_task(researcher, topic)
    writing_task = create_writing_task(writer, topic)
    editing_task = create_editing_task(editor)

    # Assemble the crew
    crew = Crew(
        agents=[researcher, writer, editor],
        tasks=[research_task, writing_task, editing_task],
        process=Process.sequential,  # Research -> Write -> Edit
        verbose=True,
        memory=True  # Enable memory for better context sharing
    )

    # Run the crew
    result = crew.kickoff()
    return result


if __name__ == "__main__":
    topic = "The impact of AI agents on software development in 2025"
    print(f"\nStarting research crew for topic: {topic}\n")
    print("=" * 60)

    final_article = run_research_crew(topic)

    print("\n" + "=" * 60)
    print("FINAL ARTICLE")
    print("=" * 60)
    print(final_article)

    # Save to file
    with open("output_article.md", "w") as f:
        f.write(str(final_article))
    print("\nArticle saved to output_article.md")

python main.py

Starting research crew for topic: The impact of AI agents on software development in 2025
============================================================

[Researcher] Starting task: Conduct thorough research...
[Researcher] Using tool: Web Search
  Query: "AI agents software development 2025 market size"
  Found 5 results...
[Researcher] Using tool: Web Search
  Query: "AI coding assistants impact developer productivity 2025"
  Found 5 results...
[Researcher] Using tool: Read Webpage
  Reading: https://example.com/ai-agents-report...
[Researcher] Task complete. Output: 1,100 words research brief.

[Writer] Starting task: Write comprehensive article...
[Writer] Task complete. Output: 1,350 words article draft.

[Editor] Starting task: Review and edit the article...
[Editor] Task complete. Output: Final polished article with editor's note.
============================================================
FINAL ARTICLE
============================================================
Editor's Note: Strengthened the opening hook with a specific statistic,
tightened transitions between sections, and clarified technical terminology...

# The Rise of AI Agents in Software Development
...

fact_checker = Agent(
    role="Fact Checker",
    goal="Verify all claims and statistics in the article against primary sources",
    backstory="""You are a rigorous fact-checker for a major news organization.
    You verify every claim against primary sources and flag anything unsubstantiated.""",
    tools=[web_search, read_webpage],
    verbose=True,
    llm="gpt-4o"
)

fact_check_task = Task(
    description="Verify all factual claims in the article. Flag any that cannot be verified.",
    expected_output="The article with [VERIFIED] or [UNVERIFIED] tags next to each claim.",
    agent=fact_checker
)

crew = Crew(
    agents=[researcher, writer, editor],
    tasks=[research_task, writing_task, editing_task],
    process=Process.hierarchical,
    manager_llm="gpt-4o"  # Manager delegates tasks dynamically
)

research_task = Task(
    description="Research the topic: {topic}",
    expected_output="A research brief on {topic}.",
    agent=researcher
)

# Pass inputs at kickoff
result = crew.kickoff(inputs={"topic": "quantum computing breakthroughs"})

research-crew/
├── .env                  # API keys (gitignored)
├── .gitignore            # Ignore .env and venv
├── agents.py             # Agent definitions
├── tasks.py              # Task definitions
├── tools.py              # Custom tools (web search, webpage reader)
├── main.py               # Orchestration and entry point
├── output_article.md     # Generated article output
└── requirements.txt      # Dependencies