Financial AI Agent
An AI agent providing unified access to financial market data and news articles.
š Model Context Protocol(MCP) Agent App
A TypeScript-based Ai agent with MCP integration. This project provides a unified interface to access financial market information and relevant news articles through MCP servers with tool calling.
DEMO
Terminal
https://github.com/user-attachments/assets/47150b98-0663-43f6-927e-f09abb796fe6
Cursor
https://github.com/user-attachments/assets/72210b49-dc23-4bef-b7ee-6d289ede59e1
⨠Features
- š Stock Market Data:
- Company Information
- Quarterly Financials
- Annual Financial Reports
- š° News Integration:
- Real-time News Articles
- Query-based News Search
- š ļø Architecture Components:
- Client-Server Architecture
- TypeScript Implementation
- Modular Design Pattern
- LangChain agent
š§ Prerequisites
- Node.js 16.x or higher
- TypeScript 4.x or higher
- API keys for stock services and OpenAi
š Quick Start
1. Installation
Clone and set up the project:
git clone https://github.com/extrawest/financial-ai-agent-mcp-servers
cd financial-ai-agent-mcp-servers
# Install dependencies
npm install
2. Environment Configuration
Create a .env
file in the project root:
# API Configuration
STOCK_API_KEY=your_stock_api_key
OPENAI_API_KEY=your_open_ai_key
3. Build and Run
# Build the project
npm run build
# Start the server
npm start
šļø Project Structure
src/
āāā agent/
ā āāā agent.ts # Agent implementation
āāā client/
ā āāā client.ts # API client implementation
āāā config/
ā āāā configs.ts # Configuration management
āāā logger/
ā āāā logger.ts # Logging functionality
āāā server/
ā āāā news-server.ts # News API server
ā āāā stocks-server.ts # Stock market API server
āāā types/
ā āāā newsapi.d.ts # Type definitions
āāā index.ts # Application entry point
š¦ Core Components
Stock Server
Handles all stock market related operations:
- Company information retrieval
- Financial reports access
- Market data processing
News Server
Manages news-related functionality:
- Article search and retrieval
- Query processing
- News data formatting
Client
Provides a unified interface to interact with both servers:
- API request handling
- Response processing
- Error management
Start an app
In separate terminal tabs run
# Start Stocks server
npm run stocks:built
# Start News server
npm npm run news:built
# Run agent
npm run agent:built
Add mcp to Cursor IDE
- Open Cursor IDE, go to Settings -> Cursor Settings
- Go to Tools & Integrations section
- Find MCP tools, click New MCP Server
- Add this to mcp.json file
{
"mcpServers": {
"stocks-mcp-server": {
"url": "http://127.0.0.1:3000/sse"
},
"news-mcp-server": {
"url": "http://127.0.0.1:3001/sse"
}
}
}
- Run commands in separte terminal tabs
# Start Stocks server
npm run stocks:built
# Start News server
npm npm run news:built
š License
This project is licensed under the MIT License - see the LICENSE file for details.
Created by Oleksandr Samoilenko
Extrawest.com, 2025
Related Servers
SearXNG MCP Server
A web search server powered by the SearXNG API.
ClinicalTrials MCP Server
Search and access clinical trial data from ClinicalTrials.gov.
Claude Web Search
Provides web search functionality powered by the Claude API from Anthropic.
Wttr Weather
Fetches weather data from the wttr.in service.
Java Conferences MCP Server
Provides information about Java conferences worldwide by parsing data from the javaconferences.github.io repository.
ęęę 大ę°ę®ęå”
Provides comprehensive bidding and tender information query services, including statistics, searches, and planned project queries.
Perplexity MCP Zerver
Interact with Perplexity.ai using Puppeteer without an API key. Requires Node.js and stores chat history locally.
MCP Documentation Server
A server for document management and semantic search using AI embeddings, with local JSON storage.
avr-docs-mcp
This MCP (Model Context Protocol) server provides integration with Wiki.JS for searching and listing pages from Agent Voice Response Wiki.JS instance.
Deep Research
An agent-based tool for web search and advanced research, including analysis of PDFs, documents, images, and YouTube transcripts.