The natural-language screen removes the step where an idea gets translated into filter fields, which is where most screeners lose people. scan_patterns runs the query in reverse: not what is AAPL doing, but who is forming a double bottom right now.
A hosted stock screener and chart-pattern detector at https://api.getcotrader.com/mcp. Screening is expressed in natural language across price, volume, moving averages, momentum, volatility and fundamentals; pattern detection covers support and resistance, trendlines, channels, double and triple bottoms, pennants and head and shoulders.
- A natural-language screen across 11,000+ US stocks — oversold on RSI while still trading above its long-term moving average, say (screen_stocks)
- Chart patterns detected for a specific stock (detect_patterns), or the entire universe scanned for one pattern type (scan_patterns)
- Ticker symbols found by name or prefix (search_symbols)
- Historical daily OHLCV price data (get_stock_bars)
- Screening criteria spanning close, open, high, low and price change over 1d to 1y; volume, average volume and relative volume; SMA and EMA over short, medium and long lookbacks; RSI, MACD and Stochastic; Bollinger Bands and ATR; and market cap, P/E ratio, sector and industry
An API key from getcotrader.com, passed either in the URL as an api_key query parameter or as an X-API-Key header. Nothing is installed locally.
One command plus a key — npx -y mcp-remote https://api.getcotrader.com/mcp, then supply credentials