Releases · goldspanlabs/optopsy-mcp

First Release

A high-performance options and stock backtesting engine exposed as an MCP server.

32 options strategies — singles, verticals, straddles, strangles, butterflies, condors, iron condors/butterflies, calendars, diagonals, covered call, protective put
Signal-driven stock backtesting on OHLCV data with 67 built-in formula functions
HMM regime filtering — gate entries by Hidden Markov Model market regime with online forward filtering (no look-ahead bias) and posterior threshold hysteresis
Cross-symbol signals — reference other symbols in formulas (e.g., VIX > 20 and rsi(close, 14) < 30)
Parameter sweep with out-of-sample validation, permutation testing, and sensitivity analysis
Walk-forward analysis with rolling train/test windows
Portfolio backtesting — run multiple strategies as a weighted portfolio with correlation matrix
Risk analytics — Monte Carlo simulation, factor attribution, drawdown analysis, cointegration testing
5 position sizing methods — fixed, fixed fractional, risk per trade, Kelly criterion, volatility targeting
4 slippage models — mid, spread (bid/ask worst-case), liquidity-based, per-leg fixed
23 MCP tools for backtesting, optimization, validation, and analysis
MCP transport — stdio (Claude Desktop) or HTTP (StreamableHTTP on any port)

git clone https://github.com/goldspanlabs/optopsy-mcp.git
cd optopsy-mcp
cargo build --release

Add to Claude Desktop config:

{
  "mcpServers": {
    "optopsy": {
      "command": "/path/to/optopsy-mcp/target/release/optopsy-mcp"
    }
  }
}

Place Parquet data in ~/.optopsy/cache/ — see README for schema details.