trading-cli

A standalone Go binary for technical analysis, screening, backtesting, sentiment analysis, and news aggregation for financial markets.

Designed to be used as a CLI tool or wrapped as an MCP (Model Context Protocol) server for AI agents. No TradingView account or authentication required — all data sourced from public APIs.

Features

• Screener Tools — Top gainers/losers, Bollinger Band scans, rating filters
• Single Asset Analysis — Full technical analysis with 23 indicator groups including RSI, MACD, Bollinger Bands, and more (see docs/TOOLS.md for full list)
• Candle Pattern Detection — Consecutive candles, advanced multi-candle patterns
• Volume Scanners — Breakout detection, confirmation analysis, smart volume with RSI
• Multi-Timeframe Analysis — Cross-timeframe alignment and divergence detection
• Sentiment & News — Reddit sentiment analysis, RSS financial news, combined confluence
• Backtesting Engine — 6 strategies (RSI, Bollinger, MACD, EMA Cross, Supertrend, Donchian) with walk-forward analysis
• Yahoo Finance — Real-time quotes and global market snapshots
• Trade Planning — Stock scoring, trade setup, Fibonacci retracement

Installation

Homebrew (macOS/Linux)

brew tap alorse/homebrew-tap
brew install trading-cli

Go install

go install github.com/alorse/trading-cli/cmd/trading-cli@latest

Binary download

Download pre-built binaries for your platform from the releases page. Available for Linux, macOS, and Windows (amd64 and arm64).

Documentation

• Tools Reference — Complete documentation for all 25 commands with flags, examples, and output formats

Usage

# Top gainers on KuCoin
trading-cli top-gainers --exchange KUCOIN --timeframe 15m --limit 10

# Top gainers on Binance Futures
trading-cli top-gainers --exchange BINANCE --timeframe 15m --limit 10 --futures

# Full analysis of Bitcoin
trading-cli coin-analysis --symbol BTCUSDT --exchange KUCOIN --timeframe 1h

# Analysis of a perpetual futures symbol
trading-cli coin-analysis --symbol BTCUSDT.P --exchange BINANCE --timeframe 1h

# Backtest RSI strategy on Apple
trading-cli backtest-strategy --symbol AAPL --strategy rsi --period 1y

# Compare all strategies on Ethereum
trading-cli compare-strategies --symbol ETH-USD --period 2y

# Reddit sentiment for Bitcoin
trading-cli market-sentiment --symbol BTC --category crypto

# Market snapshot
trading-cli market-snapshot

# Health check
trading-cli health

All commands output structured JSON to stdout. Errors go to stderr with a non-zero exit code.

MCP Server

trading-cli can run as an MCP server over stdio, exposing all 25 tools to AI agents like Claude Code, Cursor, or any MCP-compatible client.

Quick install

trading-cli mcp install                        # Claude Desktop (default)
trading-cli mcp install --client claude-code   # Claude Code
trading-cli mcp install --client cursor        # Cursor
trading-cli mcp install --client windsurf      # Windsurf
trading-cli mcp install --client codex         # OpenAI Codex CLI
trading-cli mcp install --client vscode        # VS Code Copilot
trading-cli mcp install --client gemini        # Gemini CLI
trading-cli mcp install --client amazon-q      # Amazon Q Developer
trading-cli mcp install --client zed           # Zed
trading-cli mcp install --client lm-studio     # LM Studio
trading-cli mcp install --client --list        # show all supported clients

This command writes the MCP config into your client's settings file automatically. No manual JSON editing needed.

Use --dry-run to preview the change or --force to overwrite an existing entry.

Start the server manually

trading-cli mcp

The server speaks JSON-RPC 2.0 over stdin/stdout. No flags or configuration needed — just pipe stdin and read stdout.

Manual config

If you prefer to edit the config yourself, add this to your client's MCP settings:

{
  "mcpServers": {
    "trading-cli": {
      "command": "trading-cli",
      "args": ["mcp"]
    }
  }
}

If trading-cli is installed via Homebrew or go install, the binary is on $PATH. Otherwise, use the full path to the binary.

Available tools

All 25 CLI commands are exposed as MCP tools with the same parameters (using snake_case names). See docs/TOOLS.md for parameter details.

Development

# Build
go build -o trading-cli ./cmd/trading-cli/

# Test
go test ./... -race -cover

# Lint
golangci-lint run

License

MIT