Clone Trader

Highlights

• Check positions and place paper trades with Alpaca

  Check your brokerage account and positions, place or cancel trades, and pull price history, all from Claude Code. Covers account, positions, orders, latest quotes, and historical bars saved to CSV for backtests, with no SDK to install. Paper trading by default; live orders need ALPACA_LIVE=1 and your yes for each one. Never prints your keys. Not financial advice.

  Sample run: dry run prints a paper order without any secret

• Explain why your portfolio beat or lagged its benchmark

  Explain why a portfolio beat or trailed its benchmark, and which sectors helped or hurt. Splits active return into allocation, selection, and interaction effects (Brinson-Fachler), linked across periods with Carino so they add up exactly, plus a plain-English summary for your monthly or quarterly review. Works offline from a holdings CSV. Not financial advice.

  Sample run: writes linked attribution for the two-period example: Attribution

• Backtest a trading strategy on your price data

  Test whether a trading idea would have worked on past prices, with honest fills, slippage, commissions, and no look-ahead. Compare built-in SMA crossover, RSI mean reversion, and buy-and-hold, or plug in your own strategy file, and get equity curve and trades CSVs plus CAGR, Sharpe, Sortino, max drawdown, and win rate. Reads any OHLCV CSV. Not financial advice.

  Sample run: sma-cross on the example series writes all three outputs: Sma equity

• Calculate how many shares to buy for a trade

  Work out how many shares to buy and where your stop goes, so a losing trade costs only what you planned. Size by fixed percent risk, ATR volatility stops, or capped fractional Kelly, then run a portfolio heat check that totals risk across open positions and says whether a new trade fits. A calculator that never says whether to take a trade. Not financial advice.

  Sample run: fixed-fractional sizing writes JSON

• Keep a trading journal and review it weekly

  Log every trade with its setup, thesis, emotions, and mistakes, then get a weekly review that shows what is working and what is losing you money. Groups results by setup, time of day, holding period, emotion, and plan adherence, with win rate, expectancy, profit factor, and average R. Stored in a plain CSV. Coaches your process, never your picks. Not financial advice.

  Sample run: weekly review of the example journal finds the edge and the leak: Review