Get Signal (AI) reviews closed candles and generates validated entry, stop-loss, and take-profit levels for rule-qualified ICT setups. Free Signal uses deterministic rule-based levels without an OpenAI call. Report stop-loss or take-profit outcomes so the system can filter repeated losing patterns.
Au
GOLD · XAU/USD
Twelve Data candle feed
WAIT
—
Waiting for API…
ENTRY—
STOP LOSS—
TAKE PROFIT—
1H trend:—
Range zone:—
Reason:—
Candle pattern:—
AI review:—
Click Get Signal to analyze recent candles.
Learning uses recorded outcomes when KV is configured.
₿
BITCOIN · BTC/USD
Twelve Data candle feed
WAIT
—
Waiting for API…
ENTRY—
STOP LOSS—
TAKE PROFIT—
1H trend:—
Range zone:—
Reason:—
Candle pattern:—
AI review:—
Click Get Signal to analyze recent candles.
Learning uses recorded outcomes when KV is configured.
AI-generated trade levels
Generated per scan
Click Get Signal (AI). When the ICT rule engine identifies a BUY/SELL candidate, AI reviews the recent closed 15-minute candles and proposes entry, stop-loss, and take-profit. The Worker validates direction, distance from current price, stop-risk, and 1.5R–3R reward/risk before displaying the AI levels. If checks fail, the signal is withheld. Free Signal uses rule-based levels instead.
Important: AI-generated prices are estimates based on recent candles, not guaranteed execution prices or financial advice. Use a live chart, account for spread/slippage, and paper-trade before risking money.
TradingView charts
15m
Gold · OANDA chart (display only)
Bitcoin · Coinbase chart (display only)
Backtest results
Historical estimate
Choose a backtest button above. Results include win rate and simplified profit factor estimate; they do not guarantee future results.
No backtest run yet.
Signal history
Signals captured by this browser during this session.
TIME
ASSET
SIGNAL
ENTRY
STOP
TP
AI
No signals yet.
Important: Data availability for XAU/USD and BTC/USD depends on your Twelve Data plan and symbol entitlement. AI mode uses a separate OpenAI API key to generate entry, stop-loss, and take-profit for rule-qualified candidates; the Worker validates the proposal and rejects invalid levels. Free mode uses deterministic rule-based levels. Outcome-based learning requires the Cloudflare KV binding SIGNAL_MEMORY and manually reported results. This simplified model is not a validated strategy. Backtests omit real spread, commission, slippage, execution delays and precise intrabar ordering. Paper-trade first.
API keys are not stored in this page. Keep them as Cloudflare Worker secrets.