📈 Multi-Agent Financial Risk Intelligence Platform

An AI-powered financial analytics platform that leverages Multi-Agent Systems, FastMCP, Google Gemini, Finnhub API, and Streamlit to perform intelligent portfolio analysis, financial risk assessment, market sentiment analysis, and AI-driven investment recommendations.

The platform combines quantitative financial metrics with real-time market intelligence to help investors better understand portfolio performance and potential financial risks.

🚀 Live Demo

🔗 ** Streamlit Application**

https://multi-agent-financial-risk-intelligence-platform-dp6kcy9gcgcdc.streamlit.app/

Related MCP server: Trader Journal MCP

⭐ Key Features

• 📊 Live Portfolio Valuation

• 📈 Portfolio Allocation Analysis

• 🏦 Sector Allocation Analysis

• 📉 Value at Risk (VaR)

• 📊 Sharpe Ratio Calculation

• ⚡ Live Stock Prices using Finnhub API

• 📰 Real-Time Financial News

• 😊 News Sentiment Analysis

• 🤖 AI Financial Advisor using Google Gemini

• 🔌 FastMCP Financial Server

• 📋 CSV Report Generation

• 📊 Interactive Streamlit Dashboard

• ☁️ Cloud Deployment

Dashboard Preview

Dashboard

Portfolio Analytics

![Portfolio Analytics](images/portfolio.png)

AI Financial Advisor

images/advisor.png

Risk Analytics

Add screenshot

images/risk.png

Risk Analytics

📌 Project Overview

The Multi-Agent Financial Risk Intelligence Platform is an end-to-end financial analytics application designed to simulate an intelligent financial assistant capable of analyzing investment portfolios using multiple specialized AI agents.

The system combines:

• Live Portfolio Valuation

• Portfolio Risk Analysis

• Value at Risk (VaR)

• Sharpe Ratio

• Market Sentiment Analysis

• Financial News Aggregation

• AI-powered Financial Recommendations

into a single interactive dashboard.

Unlike traditional portfolio trackers, this platform follows a Multi-Agent Architecture, where individual agents independently perform financial analysis before combining their outputs to generate intelligent recommendations.

🎯 Objectives

The primary objectives of this project are:

• Build an AI-powered financial analytics platform.

• Demonstrate Multi-Agent System architecture.

• Integrate FastMCP for reusable financial tools.

• Perform portfolio risk assessment.

• Provide AI-generated investment recommendations.

• Visualize financial metrics through an interactive dashboard.

• Deploy a production-ready Streamlit application.

✨ Highlights

• Multi-Agent Architecture

• FastMCP Financial Server

• Google Gemini Integration

• Finnhub Live Market Data

• Streamlit Interactive Dashboard

• Financial News Aggregation

• Market Sentiment Analysis

• Portfolio Allocation Visualization

• Sector Allocation Visualization

• Value at Risk (95%)

• Sharpe Ratio

• Cloud Deployment

🏗️ System Architecture

                           User
                             │
                             ▼
                  Streamlit Dashboard
                             │
     ┌───────────────────────┼────────────────────────┐
     │                       │                        │
     ▼                       ▼                        ▼
Portfolio Agent        Risk Analysis Agent      AI Advisor Agent
     │                       │                        │
     │                       │                        │
     ▼                       ▼                        ▼
Live Portfolio         VaR + Sharpe Ratio       Google Gemini
     │                       │                        │
     └───────────────┬───────────────┬────────────────┘
                     ▼
             FastMCP Financial Server
                     │
      ┌──────────────┼──────────────┐
      │              │              │
      ▼              ▼              ▼
 Finnhub API     News Agent     Recommendation
 Live Prices      RSS Feed          Engine
                     │
                     ▼
             Final Financial Insights

🤖 Multi-Agent Workflow

The application follows a modular Multi-Agent Architecture, where every agent performs a dedicated financial task before contributing to the final investment recommendation.

📊 Portfolio Agent

Responsible for:

• Reading portfolio holdings

• Calculating live portfolio value

• Computing stock allocation

• Computing sector allocation

📉 Risk Analysis Agent

Calculates important financial risk metrics including:

• Portfolio Volatility

• Value at Risk (95%)

• Sharpe Ratio

These metrics help estimate portfolio performance and downside risk.

📰 News Agent

Collects the latest financial news from reliable RSS feeds.

Responsibilities include:

• Market updates

• Company news

• Economic events

• Financial headlines

😊 Sentiment Agent

Performs sentiment analysis on financial news using TextBlob.

Each news headline is classified as:

• 🟢 Positive

• 🟡 Neutral

• 🔴 Negative

This helps estimate current market sentiment.

🤖 AI Financial Advisor

Powered by Google Gemini 2.5 Flash.

The AI combines:

• Portfolio Value

• Sharpe Ratio

• Value at Risk

• Market Sentiment

• Financial News

to answer investor questions and generate personalized financial recommendations.

🔌 FastMCP Financial Server

The project exposes reusable financial tools using FastMCP.

Available tools include:

• Portfolio Value

• Sharpe Ratio

• Value at Risk

• Latest News

• Market Sentiment

• Investment Recommendation

The MCP server enables AI models to access financial information through structured tool calls.

🛠️ Technology Stack

🔧 MCP Tools

The Financial Risk MCP Server exposes the following tools:

📂 Project Structure

multi-agent-financial-risk-intelligence-platform/
│
├── dashboard/
│   ├── app.py
│   └── components/
│       ├── advisor.py
│       ├── dashboard.py
│       ├── portfolio.py
│       ├── sidebar.py
│       └── __init__.py
│
├── Charts/
│   ├── portfolio_allocation.png
│   ├── sector_allocation.png
│   └── daily_returns.png
│
├── data/
│   └── portfolio.csv
│
├── reports/
│   └── report.csv
│
├── utils/
│   └── finance.py
│
├── llm.py
├── server.py
├── main.py
├── requirements.txt
├── README.md
└── .gitignore

⚙️ Installation

1. Clone the Repository

git clone https://github.com/mishtisethi12/multi-agent-financial-risk-intelligence-platform.git

cd multi-agent-financial-risk-intelligence-platform

2. Create a Virtual Environment

Windows

python -m venv venv

venv\Scripts\activate

macOS/Linux

python3 -m venv venv

source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

🔑 Environment Variables

Create a .env file in the project root.

GEMINI_API_KEY=YOUR_GEMINI_API_KEY

FINNHUB_API_KEY=YOUR_FINNHUB_API_KEY

Important: Never commit your .env file to GitHub.

▶️ Running the Project

Run the Analytics Engine

python main.py

Start the FastMCP Server

python server.py

Launch the Dashboard

streamlit run dashboard/app.py

☁️ Deployment

The application is deployed on Streamlit Community Cloud.

Deployment includes:

• Google Gemini API

• Finnhub API

• FastMCP Financial Server

• AI Financial Advisor

• Interactive Dashboard

To deploy:

1. Push the repository to GitHub.

2. Connect the repository with Streamlit Community Cloud.

3. Add the required API keys in Secrets:

GEMINI_API_KEY="YOUR_KEY"

FINNHUB_API_KEY="YOUR_KEY"

4. Deploy the application.

📈 Dashboard Features

The Streamlit dashboard provides:

• Live Portfolio Value

• Portfolio Allocation Chart

• Sector Allocation Chart

• Daily Returns Visualization

• Value at Risk

• Sharpe Ratio

• Market Sentiment

• AI Financial Advisor

• Financial News

• MCP Server Status

📊 Financial Metrics

Value at Risk (VaR)

Value at Risk estimates the potential portfolio loss over a specified time period at a chosen confidence level.

Confidence Level:

95%

Sharpe Ratio

Measures portfolio performance after adjusting for risk.

Higher values indicate better risk-adjusted returns.

Portfolio Allocation

Calculates each stock's contribution to the overall portfolio value.

Sector Allocation

Groups holdings into sectors to measure diversification.

Market Sentiment

Financial headlines are analyzed using TextBlob and classified into:

• Positive

• Neutral

• Negative

AI Financial Advisor

Google Gemini analyzes:

• Portfolio Value

• Portfolio Risk

• Sharpe Ratio

• VaR

• Market Sentiment

• Latest Financial News

to generate personalized investment recommendations.

📸 Screenshots

Replace these placeholders with actual screenshots after uploading them to the repository.

Dashboard

images/dashboard.png

Portfolio Analytics

images/portfolio.png

AI Financial Advisor

images/advisor.png

Risk Analytics

images/risk.png

🚀 Future Enhancements

The platform is designed with scalability in mind. Planned enhancements include:

• Multi-Portfolio Support

• User Authentication & Secure Login

• Historical Portfolio Performance Tracking

• Portfolio Optimization using Modern Portfolio Theory (MPT)

• Monte Carlo Risk Simulation

• Real-Time Market Alerts

• Interactive Stock Comparison Dashboard

• PDF Financial Report Generation

• Watchlist Management

• Docker Containerization

• CI/CD Pipeline with GitHub Actions

• Database Integration (PostgreSQL/MongoDB)

• Role-Based User Access

• Advanced Risk Metrics (Beta, Alpha, Sortino Ratio)

💡 Learning Outcomes

This project provided hands-on experience with:

• Multi-Agent System Design

• Financial Risk Analytics

• Portfolio Analysis

• FastMCP Tool Development

• Google Gemini API Integration

• Prompt Engineering

• REST API Integration

• Streamlit Dashboard Development

• Financial Data Visualization

• Git & GitHub Workflow

• Cloud Deployment

• Environment Variable Management

• Modular Python Application Development

🎯 Key Achievements

• Designed and developed an end-to-end AI-powered financial analytics platform.

• Implemented a modular Multi-Agent architecture for portfolio analysis and financial decision support.

• Built reusable financial tools using FastMCP.

• Integrated Google Gemini to provide intelligent financial insights through a conversational AI advisor.

• Connected the application with Finnhub API to retrieve live market data.

• Performed portfolio risk analysis using Value at Risk (VaR) and Sharpe Ratio.

• Visualized portfolio allocation, sector allocation, and daily returns using interactive charts.

• Deployed the application on Streamlit Community Cloud.

🌟 Why This Project?

Traditional portfolio trackers primarily display numbers without explaining investment decisions.

This project goes a step further by combining:

• Financial Analytics

• Artificial Intelligence

• Multi-Agent Systems

• Financial APIs

• Cloud Deployment

to create an intelligent financial assistant capable of delivering actionable portfolio insights.

👩‍💻 Author

Mishti Sethi

AI/ML Undergraduate

Passionate about:

• Artificial Intelligence

• Financial Analytics

• Multi-Agent Systems

• Machine Learning

• Data Analytics

• FinTech

GitHub

https://github.com/mishtisethi12

LinkedIn

Add your LinkedIn profile here

🤝 Contributing

Contributions, suggestions, and improvements are always welcome.

If you find a bug or have an idea for a new feature, feel free to open an issue or submit a pull request.

⭐ Support

If you found this project helpful or interesting, consider giving it a ⭐ on GitHub.

It really helps and motivates future development.

📄 License

This project is licensed under the MIT License.

Feel free to use, modify, and distribute this project in accordance with the license terms.

🙏 Acknowledgements

This project makes use of the following technologies and services:

• Google Gemini API

• Finnhub API

• FastMCP

• Streamlit

• Pandas

• NumPy

• Matplotlib

• TextBlob

• Python

Thank you for visiting this repository!

If you enjoyed exploring this project, don't forget to ⭐ the repository and connect with me on GitHub.