Stock Research

You can run a production-grade MCP server for stock intelligence that combines Alpha Vantage and Finnhub data, provides deterministic indicators and scoring, and delivers Claude-driven narrative analysis constrained to computed data. It supports local stdio transport for Claude Desktop usage and remote HTTP transport forRender-backed deployments, enabling flexible, scriptable stock research workflows.

How to use

You run the MCP server locally to chat with Claude via stdio, or remotely via HTTP. In stdio mode, Claude Desktop communicates with the server through a local process, while in HTTP mode you expose endpoints that your clients can call over the network. You will use the server to query stock indicators, metrics, and narratives based on the computed data and trusted data providers.

To use the server with Claude Desktop locally, ensure you set the transport mode to stdio and start the server. For remote usage, configure the transport to HTTP, supply host and port, and run so clients can connect to the endpoints you expose.

How to install

Prerequisites you need before installation: Python 3.11+, a Python virtual environment tool, and internet access to install dependencies.

Step-by-step setup commands you should run in a fresh directory:

python -m venv .venv
. .venv/Scripts/activate
pip install -r requirements.txt
copy .env.example .env

Additional setup notes

You will provide API keys for Claude, Alpha Vantage, and Finnhub. Create or update an environment file with these keys and choose a transport mode according to your deployment target.

Configuration and run guidance

Local development and runtime commands are shown in the following examples. Use the stdio transport for Claude Desktop usage and the HTTP transport for remote usage.

Claude Desktop config demonstrates how to mount the MCP server as a local process with the required environment variables.

{
  "mcpServers": {
    "stock_research": {
      "command": "python",
      "args": ["-m", "mcp_server.main"],
      "env": {
        "TRANSPORT_MODE": "stdio",
        "CLAUDE_API_KEY": "YOUR_KEY",
        "ALPHA_VANTAGE_API_KEY": "YOUR_KEY",
        "FINNHUB_API_KEY": "YOUR_KEY"
      }
    }
  }
}

Run (Claude Desktop local via stdio)

Set the transport to stdio and start the server with the following command.

# Set in your environment or within your launcher
TRANSPORT_MODE=stdio

# Start the MCP server
python -m mcp_server.main

Run (Render remote via HTTP)

For remote HTTP usage, configure the transport to HTTP and specify how you want to stream responses. Then run the server to expose endpoints suitable for SSE or streamable modes.

# Set in your environment
TRANSPORT_MODE=http
HTTP_TRANSPORT=sse  # or streamable
HOST=0.0.0.0
PORT=8000

# Start the MCP server
python -m mcp_server.main

Render deployment notes

If you deploy to Render, ensure you provide the transport and host configuration in the environment and use the MCP URL endpoints that correspond to your chosen transport mode.

Tests

Run the test suite to verify indicators, metrics, and scoring logic.

pytest -q

Available tools

stock_research_report

Primary tool for generating stock research reports using integrated data from Alpha Vantage and Finnhub with narrative analysis.

analyze_stock

Tool to analyze a stock using multiple indicators and scoring metrics.

get_price

Fetches the latest price for a given ticker.

get_ohlcv

Retrieves open/high/low/close/volume data for a symbol.

get_technicals

Obtains technical indicators such as RSI and MACD.

get_fundamentals

Gathers fundamental financial data for stocks.

get_news_sentiment

Assesses sentiment from stock-related news.

get_stock_price

Returns current price data for a symbol.

get_quote

Retrieves market quotes for a symbol.

get_company_profile

Fetches company profile details.

get_candles

Provides candlestick data for charting.

get_stock_news

Retrieves latest stock news articles.

get_rsi

Calculates the Relative Strength Index for a symbol.

get_macd

Calculates the MACD indicator.

get_key_financials

Fetches key financial statements and metrics.