Ask in plain language
Ask about price action, funding, positioning, news or macro. The answer is built from collected data, and numbers that cannot be traced to a source are not allowed through.
Open the research terminalAETHER answers market questions with the data behind them, and checks your trading ideas against real price history before you risk money. If the evidence is weak, it tells you.
Research tool, not investment advice. Past results do not predict future returns.
This is the real app, driven by a script. The backtest result is a real sandbox run saved on 2026-11-11 (6 trades), not live and not a recommendation. Open the live app →
Real BTC/USDT candles from Binance. Press ▶ to replay the last 24 hours, drag to scrub.
Real BTC/USDT 5m candles (Binance), last 24 hours.
Start with a question, or bring a strategy. Either way you see the numbers and where they come from.
Ask about price action, funding, positioning, news or macro. The answer is built from collected data, and numbers that cannot be traced to a source are not allowed through.
Open the research terminalWrite or paste Python (or describe the idea in words), run it on real candles, and replay every trade. The report shows the data period it used, and warns you when most trades exit one bar after entry.
Open the sandboxBefore anything reaches this site as a “signal”, we run it on real data against random-entry controls. Pick a family below to see what we found, including the results that make the product look worse.
RSI, MACD, moving-average ribbon, Bollinger, VWAP and more, on BTC and ETH at 1h, 4h and 1d. Round-trip costs of about 0.3% plus funding are enough to erase a 50% win rate.
R = one unit of risk. +1R means the trade earned as much as it risked.
Results from our own tests on historical Binance, Bybit and Hyperliquid data (checked 2026-10-21). They describe the past, depend on the costs we assumed, and are not a promise about the future.
Pick a market and ask what is happening and why. See the data the answer used.
Turn the idea into code or a short description and backtest it on real history.
Run it as a paper bot that places no real orders, then decide whether it deserves real money.