Newsletter Generator

from agno.agent import Agent
from agno.models.nebius import Nebius
from agno.tools.firecrawl import FirecrawlTools
from agno.storage.sqlite import SqliteStorage
import os
from dotenv import load_dotenv

load_dotenv()

# Get API keys from environment
FIRECRAWL_API_KEY = os.getenv("FIRECRAWL_API_KEY")
NEBIUS_API_KEY = os.getenv("NEBIUS_API_KEY")

# Newsletter Research Agent
newsletter_agent = Agent(
    model=Nebius(
        id="meta-llama/Meta-Llama-3.1-70B-Instruct",
        api_key=NEBIUS_API_KEY
    ),
    tools=[
        FirecrawlTools(
            search=True,
            formats=["markdown", "links"],
            search_params={
                "limit": 2,
                "tbs": "qdr:w",  # Past week
            },
        ),
    ],
    description="""You are NewsletterResearch-X, an elite research assistant 
    specializing in discovering and extracting high-quality content for 
    compelling newsletters.""",
    instructions=[
        "Use firecrawl_search to find recent articles about the topic",
        "Extract key insights, trends, and expert opinions",
        "Create compelling headlines and engaging content",
        "Follow the newsletter template structure",
        "Maintain proper attribution and citation"
    ],
    markdown=True,
    show_tool_calls=True,
    storage=SqliteStorage(
        table_name="newsletter_agent",
        db_file="tmp/newsletter_agent.db",
    )
)

def NewsletterGenerator(
    topic: str, 
    search_limit: int = 5, 
    time_range: str = "qdr:w"
) -> Dict[str, Any]:
    """
    Generate a newsletter based on the given topic and search parameters.
    
    Args:
        topic: The topic to generate the newsletter about
        search_limit: Maximum number of articles to search and analyze
        time_range: Time range for article search (e.g., "qdr:w" for past week)
    
    Returns:
        Processed newsletter content with structured metadata
    """
    try:
        # Update search parameters
        newsletter_agent.tools[0].search_params.update({
            'limit': search_limit,
            'tbs': time_range
        })
        
        response = newsletter_agent.run(topic)
        return response
    except Exception as e:
        raise RuntimeError(f'Newsletter generation failed: {e}') from e

storage=SqliteStorage(
    table_name="newsletter_agent",
    db_file="tmp/newsletter_agent.db",
)

newsletter_agent/
├── app.py              # Streamlit web interface
├── main.py             # Core agent workflow and newsletter generation logic
├── requirements.txt    # Python dependencies
├── demo.gif            # Demo animation
├── tmp/
│   └── newsletter_agent.db  # Local database for agent storage
└── README.md           # Documentation