Build rule-driven agents to test ideas, automate execution and manage risk across Binance, OKX, Bybit, Bitget and Hyperliquid with Coinrule.
410k+
users building rules
2M+
strategies executed
20+
Exchanges Supported
Instead of wiring scripts, agent based market trading becomes practical when you express behavior as clear rules. In Coinrule, each agent watches price, indicators and time, then places orders automatically on Binance, OKX and Hyperliquid.
Traders often treat an agent as a decision loop: observe, decide, act and record. You define inputs like RSI on 1-hour candles, a moving average cross, or a 24-hour volatility threshold. Then you set actions like market buy, limit sell, or reduce position size. The platform keeps an audit trail so you can review why each order fired and refine logic over time.
When you run several bots at once, coordination matters more than clever entries. Use separate agents for entries, exits and hedges, each with its own guardrails. For example, one agent can trade breakouts while another pauses new buys if total drawdown exceeds 6%. This modular approach makes changes safer because you adjust one component without rewriting everything.
Agent Design Patterns
Start with one trigger and one action, then layer filters. A simple rule might buy 2% of balance if price drops 5% within 4 hours and RSI is below 35. Add a second condition like volume above its 20-period average to avoid thin moves. Keep exits explicit: take profit at 3% and set a stop loss at 1.5% to define risk upfront.
Execution And Control
Coinrule connects via API keys so funds stay on your exchange account while automation sends orders. You can pause, clone, or schedule agents and set allocation limits per coin or per strategy. Use notifications to catch unusual behavior early, like repeated triggers during a fast market.
Grid Trading
Places staggered buy and sell limits around a range. Re-centers after a breakout. Works best on 15m to 4h charts with tight risk caps.
Volatility Breakout
Buys when price breaks above an ATR-based level on 1h candles. Exits on a trailing stop once momentum fades. Adds a cooldown to reduce whipsaws.
Stop Loss and Trailing Take Profit
Enters on your chosen signal, then manages risk automatically. Uses a fixed stop loss and a trailing take profit that tightens after gains. Built for consistent exits.
RSI Oversold Rebound
Buys when RSI drops below 30 on 4h candles. Sells when RSI recovers above 55 or price hits a profit target. Adds a time stop to avoid dead money.
MA Crossover with Volume Filter
Trades a fast and slow moving average cross on 1h charts. Confirms entries only when volume is above its 20-period average. Exits on the opposite cross or a stop loss.
Grid Trading
Places staggered buy and sell limits around a range. Re-centers after a breakout. Works best on 15m to 4h charts with tight risk caps.
Volatility Breakout
Buys when price breaks above an ATR-based level on 1h candles. Exits on a trailing stop once momentum fades. Adds a cooldown to reduce whipsaws.
Stop Loss and Trailing Take Profit
Enters on your chosen signal, then manages risk automatically. Uses a fixed stop loss and a trailing take profit that tightens after gains. Built for consistent exits.
RSI Oversold Rebound
Buys when RSI drops below 30 on 4h candles. Sells when RSI recovers above 55 or price hits a profit target. Adds a time stop to avoid dead money.
MA Crossover with Volume Filter
Trades a fast and slow moving average cross on 1h charts. Confirms entries only when volume is above its 20-period average. Exits on the opposite cross or a stop loss.
Breakout + Retest
Waits for a breakout, then buys the retest of the level. Uses an invalidation stop below support. Targets a 2R take profit for disciplined risk-reward.
Bollinger Band Mean Reversion
Buys near the lower band when price deviates from the midline. Sells at the basis or upper band depending on strength. Best in range-bound conditions.
Supertrend Momentum
Follows trend direction using Supertrend on 4h candles. Enters on a flip to bullish and exits on the next bearish flip. Adds a max-loss guardrail per trade.
DCA Accumulator
Splits buys into scheduled tranches to smooth entry price. For example, invests $500 monthly as four $125 weekly buys. Can pause automatically during high volatility.
Portfolio Rebalance
Rebalances allocations back to targets on a schedule. Sells overweight assets and buys underweight ones. Helps control concentration risk during strong trends.
Unlike ad-hoc manual trading, systematic testing shows how a rule behaves across regimes. Run a backtest, then inspect trades to see if entries cluster around news spikes or low-liquidity hours. Try a clear experiment: trade only when ADX is above 20, or only during the London-New York overlap. If results degrade, simplify the logic and retest rather than stacking indicators.
Every bot needs boundaries that override its own signals. Set max position size per asset, a daily loss limit and a rule that disables new entries after three consecutive stop outs. With agent based market trading, you can also pause buys when 1-hour ATR rises above 2.5% of price. These controls reduce the chance that a fast market turns a small bug into a large loss.
For multi-exchange execution, keep strategies consistent across accounts. Run spot rules on Kraken while testing perps logic on Hyperliquid, or split liquidity between Binance and Bybit. The interface keeps rules and logs in one place, so you do not lose context when you switch venues. For a deeper overview of automation features, see /automated-trading.
Once one strategy is stable, add a second agent with a different job. A common setup is an entry agent, an exit agent and a hedge agent that reduces exposure when funding turns extreme. Keep them independent, but share the same portfolio limits. This makes the system easier to debug when performance changes.
Launch BotFAQ
Lars H.
Priya K.
Marco P.
Amara N.
James W.
Yuki T.
When markets trend, momentum systems shine, but mean reversion can bleed. When markets range, the opposite happens. Use a regime filter like ADX, ATR expansion, or a higher timeframe moving average to decide which agent is allowed to trade. This keeps automation aligned with conditions instead of forcing one style everywhere.
Good agents specify order types, sizing and timing. Decide whether you use limit orders to reduce slippage or market orders for certainty. Add a rule that prevents repeated entries, such as one trade per asset per 6 hours. These details often matter more than the indicator choice.
Set notifications for events that deserve attention, not every trade. Useful alerts include drawdown thresholds, sudden exposure changes, or a failed order due to exchange constraints. Review them weekly and adjust thresholds so you only get actionable signals. That keeps you informed without turning automation into noise.
If you prefer decentralized execution, you can run strategies on-chain using smart accounts. Session keys can limit permissions so an agent can trade but not withdraw. Routing can seek best liquidity across pools to reduce price impact. Keep gas and slippage settings conservative until results are consistent.
Treat every change as a hypothesis. Record what you changed, why you changed it and what metric you expect to improve, such as win rate, max drawdown, or average hold time. Compare results over the same dates to avoid cherry-picking. Over time, you build a library of behaviors that fit your risk tolerance.
Pick one market, one timeframe and one exit plan, then deploy with small size. After 30 to 50 trades, review the log and adjust only one parameter. Repeat until the behavior matches your intent. Coinrule keeps the iteration loop fast, so you spend less time clicking and more time improving the system.
Start Free
