Content Strategy Crew by KSRX

A powerful AI-powered content strategy system built with crewAI that orchestrates multiple specialized AI agents to analyze data, research trends, and develop comprehensive content strategies.

Overview

The Content Strategy Crew automates the entire content planning process through a collaborative workflow where data flows from specialized agents to a central strategist:

graph TD
    subgraph "Content Strategy Crew"
        A[Crew Initialization]

        subgraph "Data Collection & Analysis"
            B[Historical Data Analyst]
            C[Company Background Analyst]
            D[Data Provider Agent]
        end

        subgraph "Strategy Development"
            E[Content Strategist]
        end

        A --> B
        A --> C
        A --> D

        B -->|Executes| F[Social Media Analysis Task]
        C -->|Executes| G[Company Data Scraping Task]
        D -->|Executes| H[Web Search & Scraping Tasks]

        F -->|Outputs| I[analysis_report.md]
        G -->|Outputs| J[company_data_scraping_report.md]
        H -->|Outputs| K[websearch_report.md]

        I -->|Context Input| L[Content Strategy Planning Task]
        J -->|Context Input| L
        K -->|Context Input| L

        E -->|Executes| L
        L -->|Outputs| M[content_strategy_plan.md]
    end

Loading

How It Works

The system follows a sequential process where:

1. The Historical Data Analyst examines past social media performance.
2. The Company Background Analyst researches your company's profile.
3. The Data Provider Agent gathers current industry trends.
4. The Content Strategist synthesizes all this information to create a comprehensive strategy.

All data from the first three agents flows into the Content Strategist, who then creates a unified content strategy plan based on this collective intelligence.

Installation

Ensure you have Python >=3.10 <3.13 installed on your system.

# Install dependencies
pip install -r requirements.txt

# Or using UV
pip install uv
crewai install

Configuration

1. Add your API keys to the .env file:

   OPENAI_API_KEY=your_openai_api_key
   SERPER_API_KEY=your_serper_api_key
   AGENTOPS_API_KEY=your_agentops_api_key
   MONGODB_URI=your_mongodb_uri
   

2. Prepare your data:

   • Place your social media analytics CSV file at data1.csv

3. Customize agent and task configurations in the YAML files:

   • config/agents.yaml: Define agent roles and goals
   • config/tasks.yaml: Customize task descriptions and outputs

Usage

You can run the Content Strategy Crew in two ways:

1. Using the run script:

   ./run.sh

   This script will create a virtual environment, install dependencies, and start the FastAPI application.

2. Access the frontend:

   Open your browser and navigate to http://localhost:8000/. You should see the frontend.

3. Interact with the API:

   Fill in the form fields and click "Execute Crew". The output from the API will be displayed in the outputContent section, and real-time updates will be shown in the Updates section.

The system will generate output files in the ./output/ directory:

• analysis_report.md: Social media performance analysis
• company_data_scraping_report.md: Company background research
• websearch_report.md: Industry trends and competitor analysis
• content_strategy_plan.md: Final comprehensive content strategy

API Endpoints

The FastAPI application provides the following endpoints:

• POST /execute_crew: Execute the Content Strategy Crew
• GET /health: Check the health status of the application and MongoDB connection
• GET /history: Retrieve all execution history entries
• GET /history/{session_id}: Retrieve a specific execution history entry
• DELETE /history/{session_id}: Delete a specific execution history entry
• GET /: Serves the main HTML file for the frontend

Real-time Updates

The frontend displays real-time updates on which agent is running which task. This allows you to monitor the progress of the crew execution.

Displaying Markdown Output

The frontend uses marked.js to parse the Markdown content and render it as HTML, providing a formatted and readable output.

Next.js Frontend

A modern Next.js frontend is also available in the ccagent-frontend directory. To use it:

cd ccagent-frontend
npm install
npm run dev

Then navigate to http://localhost:3000 to access the improved user interface with tabs for updates and output, plus a history view of previous executions.

Key Implementation Note

To ensure the Content Strategist receives all data from previous tasks, make sure to configure the context properly:

@task
def content_strategy_planning(self) -> Task:
    return Task(
        config=self.tasks_config['content_strategy_planning'],
        context=[
            self.social_media_analysis(),
            self.company_data_scraping(),
            self.website_scraping()
        ],
        output_file='./output/content_strategy_plan.md'
    )

This ensures that all the outputs from previous tasks flow into the Content Strategist, allowing for a truly integrated strategy.

Process Flow

sequenceDiagram
    participant Crew as Content Strategy Crew
    participant HDA as Historical Data Analyst
    participant CBA as Company Background Analyst
    participant DPA as Data Provider Agent
    participant CS as Content Strategist

    Crew->>HDA: Analyze social media data
    HDA->>Crew: Return performance analysis
    Crew->>CBA: Research company background
    CBA->>Crew: Return company insights
    Crew->>DPA: Search & scrape industry trends
    DPA->>Crew: Return research findings
    Crew->>CS: Synthesize all data
    Note right of CS: Receives context from all previous tasks
    CS->>Crew: Deliver comprehensive strategy

Loading

MongoDB Integration

The application now includes MongoDB integration for storing execution logs and outputs. Make sure to set the MONGODB_URI in your .env file.

CORS Support

The API now supports Cross-Origin Resource Sharing (CORS) to allow requests from the frontend.

Creator

This project was created by Kaviru Hapuarachchi.

• Website: kaviru.cc
• Email: hello@kaviru.cc
• GitHub: @Kavirubc