Are the signals generated by AI?

No. They are produced by five fixed technical rules that are printed on every ticket. A language model is used only to explain, in words, the figures those rules already computed - it is forbidden from inventing a number or predicting a price.

How often are signals updated?

The scanner re-scores every symbol on a board every few minutes, and the browser is told the moment a new one appears. You can turn on desktop notifications for them.

Which timeframes are covered?

15-minute, 1-hour, 4-hour, daily and monthly. The horizon matters more than most people expect - on futures, the same rules measure negative on 15-minute bars and positive on daily ones - so the ticket always shows which horizon it was computed on.

Can I trade a signal automatically?

Yes, on Binance spot, Binance futures and OANDA forex, through the automation engine. It will only arm a strategy and horizon whose backtests clear the evidence gate.

Crypto Trading Signals with Entry, Stop Loss & Take Profit

Rule-based crypto signals that show the entry, the stop, two targets, and every rule that fired - and that have been backtested with the same code that produces them. No black box, no "AI predicts".

A signal that says "BUY" and nothing else is an opinion. A signal that says buy at this price, stop here, take profit there, because these four rules fired and here is what each one saw is something you can check, disagree with, and measure afterwards.

Every NusaTerminal signal is the second kind. Here is exactly how one is made.

How NusaTerminal generates signals

1. Market scanning

Every few minutes the scanner re-scores every symbol on the board you are watching - twenty crypto pairs, the IDX large caps, the US names, or the forex and commodity list - on the horizon you have chosen. It does not wait for you to open a chart; a new signal is pushed to the panel and, if you have allowed it, to a desktop notification.

On the crypto board, spot and futures are scanned separately. They share a symbol and almost a price, but the perpetual has its own candles, its own funding cost and a short side that spot does not have, so a futures signal is computed on the perpetual's own bars and a spot signal on the spot pair's.

2. Technical rules

Five independent rules, each voting -1, 0 or +1:

Trend - EMA20 above or below EMA50
Position against trend - price above or below EMA20
Momentum - MACD histogram positive and growing, or negative and growing
RSI - below 30 or above 55 for a long; above 70 or below 45 for a short
Structure - price in the bottom 10% of its 40-bar range for a long, the top 10% for a short

The votes are summed. A signal is called only when the absolute score reaches 4 of 5. Three rules agreeing is shown on the symbol's own page as a weak reading; it never reaches the scanner, because a notification has to earn the interruption.

3. Trend confirmation

The first two rules are the confirmation: a long needs the short average above the long one and the price above the short one. A price bouncing in a downtrend can light up momentum and RSI and still not reach four, because the trend rules are voting against it. That is deliberate - the rules are designed so that the highest possible score in a downtrend is lower than the threshold.

4. Entry

The entry is the last price at the moment the rules fired. In the backtest, the trade is entered at the open of the next bar, never at a close that could not have been known at the time. The live ticket and the backtest use the same convention, so the numbers are comparable.

5. Stop loss

1.5× ATR(14) from the entry. ATR is the average true range - measured volatility over the last fourteen bars - so the stop is wider on a volatile pair and tighter on a quiet one, rather than a fixed percentage that is too tight for one and too loose for the other.

6. TP1

1.5× ATR from the entry, in the direction of the trade. Risk to reward is 1:1 at the first target.

7. TP2

3× ATR from the entry. Risk to reward is 1:2. When a single bar touches both the stop and a target, the backtest counts the stop as hit first - the order inside a bar cannot be known from OHLC, so the pessimistic assumption is taken.

8. Backtesting

Every signal type, on every market and horizon, is replayed over years of history using this exact code: the same rules, the same entry convention, the same stop and targets, the same tie-break, and the real cost model for that market - taker fees, slippage, and funding on futures.

The result is not just an average. It is a lower bound - the worst figure the data is still consistent with, given how many trades it rests on. A strategy with a spectacular average on forty trades and a lower bound below zero is not run by the automation engine, and the signal panel tells you so.

How the backtesting works →

What a ticket looks like


SHORT · FUTURES · BTCUSDT · 4H
score −4 of 5 rules firing · confidence 80%

entry      78,488.50
stop loss  79,596.49    1.5×ATR · risk 1,107.99
TP1        77,380.51    RR 1:1
TP2        76,272.52    RR 1:2

funding +0.03%/day · received by this side

▼ Trend             EMA20 below EMA50
▼ Position          price below EMA20
▼ Momentum          MACD histogram negative and growing
▼ RSI               41.3 - below 45
•  Structure         price mid-range - neutral

Every number on it was computed, and every rule says what it saw. If you would have called it differently, you can see exactly where you disagree.

What the signals do not do

They do not predict. They report that a fixed set of conditions is currently true, and they attach the exits that the backtests were measured with. Whether those conditions have led anywhere in the past is a separate question, and the answer is on the backtesting page - per strategy, per market, per horizon, with the sample size beside every figure.

We don't promise profits. We test the strategy first.