FAQ

No. You connect your own exchange keys with trade-only permissions; we send orders to your account. We never have withdraw permission. We never custody money.

Encrypted at rest with a per-user key. Decrypted only in memory when your strategy is running. Never logged, never sent to third parties. Rotate or delete from the profile page anytime.

No. The rule engine is a visual designer + YAML. Custom indicator plugins (Python) are a Desk-plan add-on; most users never need them.

The walk-forward optimizer is built to catch that. Robustness score under 0.4 means the strategy is curve-fit — don't paper-trade it. The Monte-Carlo and stress-test tools are designed to surface the worst-case before you risk a cent.

Alpaca for US equities and crypto, plus any CCXT-supported venue: Binance, Kraken, Bybit, OKX, Coinbase, Bitfinex — over 100 in total. Paper trading works without an exchange key, against synthetic or read-only-key data.

Decisions are made on bar close. For 1-minute bars that's roughly a one-second decision-to-order window. The engine is not a high-frequency tool — you write strategies that survive across hours, not microseconds.

On Desk plans, yes. Drop a Python file implementing the _Indicator interface (see src/quantforge/strategies/custom/indicators.py for the contract) and register it. Free and Pro plans use the built-in catalogue only.

The dashboard talks to a FastAPI server. Personal-use API access is available on Desk plans — endpoints for running backtests, listing strategies, and reading runs. Email us if you need a token.

On our servers in the EU/EEA, encrypted at rest and in transit. We do not sell or share your data. Export everything as JSON anytime from /api/account/export. Delete your account in one click.

Email security@noonbarbari.xyz. We acknowledge within 48 hours. Please don't open a public issue for vulnerabilities.

A six-indicator family that detects price-action patterns from public ICT/SMC literature: Break of Structure, Change of Character, Fair Value Gaps, Order Blocks, Equal Highs/Lows, and Premium/Discount zones. All clean-room implementations — we wrote them from scratch from the public algorithms, not from any third-party source.

SuperTrend (the built-in `ta.supertrend`), Blackflag FTS (Jose Azcarate's variant), Trend Magic (KivancOzbilgic's CCI-gated ATR ratchet), and LazyBear's Volume Flow Indicator (VFI). All four ports are in our indicator registry alongside the SMC family and the classic indicators (EMA, RSI, MACD, Bollinger, ADX, Stochastic, ATR, VWAP).

After a backtest, every closed position is drawn as a green TP zone (the planned profit target band) and a red SL zone (the planned stop band) overlaid on the candles. The zone that "won" the trade is bold; the losing side fades to background. Hover any box for a tooltip with side, entry/exit price, P&L, and qty.

Yes — every backtest-able rule set can also be deployed to paper trading via the "Run a test" page (set mode = paper). For live trading, a paper-tested setup can be promoted via the Telegram bot's deploy command. Always run paper for at least 2-4 weeks of bars before considering real capital.

Two paths. (1) In the Strategy Designer, edit the indicator's parameters directly (e.g. change `swing_length` from 50 to 20 on `market_structure`). (2) For more rigorous tuning, use the Auto-Tune feature on the Run a test page — it sweeps a parameter range and picks the best by your chosen metric. For robust tuning, follow Auto-Tune with a walk-forward validation rather than picking the best single-window result.

We did not look at or translate any specific platform's existing SMC code (e.g. LuxAlgo's CC BY-NC-SA Pine script). We implemented the underlying concepts — which are decades-old public price-action ideas from ICT educational content — from textbook descriptions. The result is a license-clean implementation we own.

Most often: (a) the indicators haven't warmed up enough by the time the test starts — try increasing `days` to fetch more bars; (b) the rule uses `crosses_*` with `{bar: close}` on the LHS — wrap close in an SMA(1) indicator since `{bar: close}` collapses prev==now in the evaluator; (c) parameters are too restrictive — try lowering thresholds.

Pick any two strategies from the library and the Compare page shows them side-by-side: indicator stack, entry conditions, exit conditions, risk profile, expected behavior. Useful when deciding which template to clone for your own variant.