I built a trading intelligence MCP server in 2 days — here's how

Dev.to / 4/16/2026

💬 OpinionDeveloper Stack & InfrastructureSignals & Early TrendsTools & Practical Usage

共有:

Key Points

The author built “Zivio MCP,” a trading-intelligence MCP server that exposes multiple market and regulatory data tools for use by AI agents.
The server provides five MCP tools: gap_scan for pre-market gappers, sec_alert for live SEC filings with Gemini AI summaries, news_sentiment for headline polarity scoring, insider_scan for insider purchases, and momentum_alerts for trading signals like squeezes and VWAP reclaim.
Users can connect the MCP server via Cursor or Claude Desktop by adding a configuration entry (including an API key header) in the respective local JSON config.
The implementation uses a Node.js/Express backend on Railway, PostgreSQL for storing API keys, Stripe webhooks for provisioning, Brevo for email delivery, and external data/AI services including Polygon, sec-api.io, Benzinga, and Gemini APIs.
The project is offered as a paid service ($49 CAD/month) with links to Stripe checkout and a GitHub repository for the codebase, inviting questions about the build.

I've been day trading gap momentum stocks for years. I built NikTrader as my personal dashboard using Polygon, Benzinga, and sec-api.io.

Last week I packaged that entire data stack as an MCP server — Zivio MCP — so any AI agent can call it directly.

What it does

5 tools available via MCP:

gap_scan — pre-market gappers with price, gap%, volume, float
sec_alert — live SEC 8-K/S-1 filings with Gemini AI summary
news_sentiment — bullish/bearish scoring per headline
insider_scan — insider purchases sorted by transaction value
momentum_alerts — squeeze, HOD, VWAP reclaim, vol spike

How to connect

Add to your .cursor/mcp.json or Claude Desktop config:

{
  "mcpServers": {
    "zivio-trading": {
      "url": "https://zivio-mcp-production.up.railway.app/mcp",
      "headers": {
        "x-api-key": "YOUR_API_KEY"
      }
    }
  }
}