🧠 AI-Powered Financial Content Creator

A multi-agent AI system that autonomously creates, refines, and formats high-quality financial content — including blog posts and social media snippets — using real-time market data and news.

---

🚀 What It Does

This system uses CrewAI agents to automate the entire content creation pipeline:

1. 📡 Market News Monitor Agent
   Tracks the latest financial news based on a given subject (e.g. “US-China tariffs”), summarizing impactful headlines.

2. 📊 Data Analyst Agent
   Analyzes market trends and economic indicators to identify actionable insights.

3. ✍️ Content Creator Agent
   Generates engaging blog content and social media posts from insights provided by the above agents.

4. 🧐 Quality Assurance Agent
   Refines and formats the content using markdown, ensuring clarity, structure, and brand alignment.

---

💼 Use Case

Ideal for:

• Financial blogs and media platforms
• Automated newsletter creation
• Market intelligence publishing
• Anyone needing timely, data-driven financial content

---

🛠️ Tech Stack

• CrewAI – Multi-agent orchestration
• LangChain / LLMs – Natural language generation
• YAML Configs – Agent & task modular setup
• Jupyter Notebook – For testing and orchestration
• Markdown output – For clean blog-ready formatting

---

🧪 Performance Benchmarks

Agent	Task Description	Avg Time (approx)
📡 Market News Monitor	Scrapes live financial news	~45 seconds
📊 Data Analyst Agent	Extracts and interprets insights	~60 seconds
✍️ Content Creator Agent	Generates blog + social media content	~75 seconds
✅ Quality Assurance Agent	Formats content in markdown	~30 seconds
⏱️ Total Runtime	End-to-end pipeline execution	~3.5 minutes

🧠 Benchmarks based on mistral-large-latest model and real-time web scraping.

---

🤖 LLM Models Used

This project uses open-weight models from Mistral.

🔹 mistral-small-latest

• Lightweight and fast
• Best for retrieval tasks and embeddings
• Used in: WebsiteSearchTool, embedder, lightweight analysis

🔹 mistral-large-latest

• Strong reasoning and summarization capabilities
• Handles long, structured generation tasks
• Used in: All core agents (llm=llms['large']) — blog writing, analysis, formatting

---

🧪 How to Run

1. Install dependencies:

pip install -r requirements.txt

2. Set your environment variables:

export MISTRAL_API_KEY=your_key_here
export SERPER_API_KEY=your_key_here

3. Launch the notebook:

jupyter notebook main.ipynb

4. Provide your topic (e.g. "US-China tariffs") and let the agents do the rest.