News Analysis Agent MCP Server

mcp-news-analysis-agent

A comprehensive Model Context Protocol (MCP) implementation for intelligent news analysis, featuring advanced LLM-powered sentiment analysis, AI summarization, and natural language query processing using Mistral AI.

🧠 LLM-Powered Architecture

This project leverages Mistral AI for advanced natural language processing capabilities:

• Sentiment Analysis: Uses Mistral's LLM for nuanced sentiment understanding with confidence scoring
• Text Summarization: AI-powered content summarization with customizable length and style
• Intent Detection: Smart query interpretation for natural language interaction
• Structured Outputs: JSON-formatted responses with detailed reasoning and metadata

🚀 Features

• News Fetching: Retrieve real-time news articles from RapidAPI
• Sentiment Analysis: Advanced sentiment analysis using Mistral AI with confidence scoring and detailed reasoning
• Text Summarization: AI-powered summarization using Mistral AI
• Intelligent Agent: Natural language query processing with enhanced intent detection
• MCP Architecture: Fully compliant with Model Context Protocol standards

📋 Project Structure

MCPDemo/
├── server/
│   └── mcp_server.py          # MCP server implementation
├── client/
│   └── mcp_client.py          # Enhanced MCP client with intelligent intent detection and quote parsing
├── tools/
│   ├── news_tool.py           # News fetching from RapidAPI
│   ├── sentiment_tool.py      # sentiment analysis tools
│   ├── summary_tool.py        # Text summarization tools
│   └── __init__.py
├── config/
│   ├── settings.py            # Configuration management
│   ├── .env                   # Environment variables
│   └── __init__.py
├── requirements.txt           # Python dependencies
└── README.md                  # This file

⚙️ Setup Instructions

1. Environment Setup

First, create a Python virtual environment:

# Navigate to project directory
cd "C:\Users\mayssen\Desktop\mcp project\MCPDemo"

# Create virtual environment
python -m venv .venv

# Activate virtual environment
.\.venv\Scripts\Activate  # Windows
# source .venv/bin/activate  # macOS/Linux

# Install dependencies
pip install -r requirements.txt

2. API Keys Configuration

Edit the config/.env file and add your API keys:

# News API key from RapidAPI (already provided)
RAPIDAPI_KEY=6d35e9aa82msh4c8550ffb3e08b4p15bf78jsna3f5a47eeb4d
RAPIDAPI_HOST=real-time-news-data.p.rapidapi.com

# Get your Mistral AI API key from https://console.mistral.ai/
MISTRAL_API_KEY=your_mistral_api_key_here

# Optional: Adjust other settings as needed
MAX_NEWS_ARTICLES=10
MAX_SUMMARY_LENGTH=500
LOG_LEVEL=INFO

3. Install Additional Dependencies

Some packages might need special installation:

# Ensure Mistral AI client is properly installed
pip install mistralai

# If you encounter any import errors, install packages individually:
pip install httpx langchain-mistralai fastmcp

🔧 Usage

Running the MCP Server

# Make sure you're in the project directory with activated virtual environment
python server/mcp_server.py

Running the Interactive Agent

In a separate terminal:

# Activate the same virtual environment
.\.venv\Scripts\Activate

# Run the client
python client/mcp_client.py

Example Queries

Once the agent is running, try these natural language queries:

- "Get latest news about technology"
- "Analyze sentiment of recent news about climate change"
- "Summarize news about the economy"
- "Show me top 5 news from UK"
- "How do people feel about the latest political news?"
- "Get French news about sports and analyze sentiment"
- "This new AI technology is amazing but also quite expensive" (direct text analysis)

🛠 Available Tools

1. News Tool

• Function: fetch_news
• Purpose: Fetches news articles from RapidAPI
• Parameters: topic, country, language, limit
• Example: Retrieve tech news from US in English

2. Sentiment Analysis Tool

• Function: analyze_sentiment
• Purpose: Analyzes sentiment using advanced Mistral AI LLM with confidence scoring and detailed reasoning
• Parameters: text, analysis_type (simple/detailed)
• Features:
  • Structured JSON responses with confidence scores
  • Detailed reasoning and emotion detection
  • Support for complex, nuanced sentiment analysis
  • Direct text analysis through quotes
• Example: Determine sentiment with confidence: "Mixed sentiment (0.80 confidence) - expresses both excitement and concern"

3. Summary Tool (Requires Mistral AI)

• Function: summarize_text
• Purpose: Summarizes text using Mistral AI
• Parameters: text, max_length, summary_type
• Example: Create brief summaries of long articles

4. Combined Workflows

• Function: analyze_news_sentiment
• Purpose: Fetches news and analyzes sentiment in one step