Trading MCP

What netanelavr/trading-mcp does

The netanelavr/trading-mcp MCP server provides stock research and trading-analysis tools through the Model Context Protocol. It covers technical and fundamental screening, valuation comparisons, financial health scoring, insider transactions, options sentiment, Reddit activity, and news context.

The server is relevant to the Finance & Fintech category for workflows that need structured research around one ticker or a group of stocks. Its comprehensive analysis tool combines several available data sources, including fundamental metrics, insider activity, options sentiment, news, and optionally social sentiment.

How it works

The netanelavr/trading-mcp MCP server runs locally as a compiled Node.js process. An MCP client starts dist/server.js and communicates with the exposed tools through the client’s MCP configuration. The implementation uses adapters for Finviz screening data, Barchart options data, Reddit, and OpenAI.

Stock screening accepts Finviz-style filter parameters and can return a limited set of matching securities. Fundamental tools retrieve metrics such as valuation, profitability, leverage, liquidity, efficiency, and growth indicators. Other tools analyze insider transactions over a selected period, retrieve put/call ratios by expiration, or combine multiple analyses for a ticker.

OpenAI is used for news analysis and Reddit sentiment analysis. Reddit-backed tools require Reddit API credentials, while news analysis requires an OpenAI API key. Finviz and Barchart access rely on web scraping, so rate limits or website changes may affect results.

Setup and configuration

Clone the repository, install its npm dependencies, and build the TypeScript project:

git clone https://github.com/netanelavr/trading-mcp.git
cd trading-mcp
npm install
npm run build

Configure the compiled server in an MCP-compatible client by using node as the command and the path to dist/server.js as its argument. The example configuration uses these environment variables:

• OPENAI_API_KEY for news analysis and AI-based Reddit sentiment analysis.
• REDDIT_CLIENT_ID, REDDIT_CLIENT_SECRET, REDDIT_USERNAME, and REDDIT_PASSWORD for Reddit research.

Reddit credentials can be created through Reddit’s app preferences by registering a script application. The README does not identify additional configuration variables or provide a separate database setup.

Tools and capabilities

The netanelavr/trading-mcp MCP server exposes tools for:

• Screening stocks with technical patterns, fundamental filters, ordering, and result limits.
• Retrieving financial metrics and comparing valuation measures such as P/E, forward P/E, PEG, and price-to-book.
• Calculating a 0–100 financial health score with configurable weights for profitability, liquidity, leverage, efficiency, and growth.
• Reviewing insider transactions with filters for transaction type, date range, and minimum value.
• Reading put/call ratios and associated options-market sentiment.
• Finding trending stocks across selected Reddit communities and analyzing Reddit posts or comments for ticker sentiment.
• Analyzing recent news, sentiment, market impact, and sector context.
• Running comprehensive analysis for a ticker across the supported research areas.

Limitations and notes

Reddit and OpenAI-dependent tools are not fully available without their corresponding credentials. The README specifically marks social research as requiring Reddit configuration and news analysis as requiring OpenAI configuration.

Finviz and Barchart data collection uses web scraping. The project notes that rate limiting and site changes can cause either source to fail occasionally. Reddit API rate limits may also affect high-volume requests. Results should therefore be treated as research output rather than a guarantee of trading outcomes.