Design an LLM trading agent workflow with Coinrule. Turn intent into rule-based crypto execution with logs, risk controls and multi-exchange support in minutes.
410k+
users building automations
2M+
rules executed monthly
20+
Exchanges Supported
Instead of trusting a chat response to click buttons, an llm trading agent on Coinrule becomes explicit conditions you can test and audit. Example: buy 1% when RSI(1H) drops below 30 and price is above MA200, then take profit at 3% and stop at 1.5%. Run the same logic on Binance, Kraken, Coinbase or Hyperliquid without rewriting it for each venue.
Traders use language models best as planners, not as unchecked executors. Coinrule translates intent into deterministic triggers, position sizing and exits, so behavior stays consistent. Require confirmations like volume above its 20-period average or a candle close beyond a level before any order is sent.
Every action is logged with the exact rule path that fired, the price used and the order outcome. Review why a trade happened, then refine thresholds without rebuilding your approach. For a tighter loop, schedule checks every 5 minutes and only allow entries during a defined session window.
Guardrails For AI Workflows
Set max allocation per coin, daily loss limits and cooldown timers after a stop. Example: if drawdown hits 2% in 24 hours, pause all strategies for 6 hours and send an alert. These controls keep automation predictable when conditions shift fast.
Non-Custodial By Design
Coinrule connects to exchanges via API keys, so funds stay in your account while the engine executes your rules. For on-chain automation, smart accounts and session keys can limit permissions to defined actions and time windows.
Volatility Breakout
Enter on a 4H break above the prior range high with ATR expansion. Exit on a close back inside the range or a trailing stop. Built for early trend ignition.
RSI Oversold Rebound
Buy when RSI(1H) drops below 30 and price holds above a higher-timeframe MA. Sell when RSI recovers above 55 or at a 2-4% take profit. Simple rebound logic.
Grid Trading
Place a ladder of buys and sells inside a defined range with equal spacing. Recycle inventory as price oscillates. Add a range-break stop to limit runaway risk.
Stop Loss and Trailing Take Profit
Add a hard stop and a trailing take profit to any entry signal. Example: stop at 1.8%, start trailing after 2.5% profit with a 1% trail. Helps lock gains while capping downside.
MA Crossover with Volume Filter
Buy when a fast MA crosses above a slow MA and volume is above its 20-period average. Sell on the cross back down or when momentum fades. Aims to reduce low-volume whipsaws.
Volatility Breakout
Enter on a 4H break above the prior range high with ATR expansion. Exit on a close back inside the range or a trailing stop. Built for early trend ignition.
RSI Oversold Rebound
Buy when RSI(1H) drops below 30 and price holds above a higher-timeframe MA. Sell when RSI recovers above 55 or at a 2-4% take profit. Simple rebound logic.
Grid Trading
Place a ladder of buys and sells inside a defined range with equal spacing. Recycle inventory as price oscillates. Add a range-break stop to limit runaway risk.
Stop Loss and Trailing Take Profit
Add a hard stop and a trailing take profit to any entry signal. Example: stop at 1.8%, start trailing after 2.5% profit with a 1% trail. Helps lock gains while capping downside.
MA Crossover with Volume Filter
Buy when a fast MA crosses above a slow MA and volume is above its 20-period average. Sell on the cross back down or when momentum fades. Aims to reduce low-volume whipsaws.
Bollinger Band Mean Reversion
Buy when price closes below the lower band and RSI confirms weakness. Exit at the middle band or when RSI normalizes. Best for range-bound regimes with defined volatility.
Perps Trend Follower
Open a position when 4H trend signals align and funding is not extreme. Use a time stop if price fails to progress. Designed for perpetual markets with strict risk controls.
Portfolio Rebalance
Rebalance back to target weights weekly or when drift exceeds a threshold. Reduce concentration after large moves. Useful for long-term baskets of major assets.
Opening Range Breakout
Define an early-session range and enter on a break with confirmation. Exit with a fixed stop and partial take profits. Built for structured intraday volatility windows.
Dynamic Grid with ATR Spacing
Build a grid where spacing adapts to ATR, widening in high volatility and tightening in calm markets. Add a range invalidation stop beyond key levels. Targets steadier fill behavior.
For centralized venues, link accounts on Binance, OKX, Bybit, Bitget or Coinbase and run the same rule set across them. Route different strategies to different exchanges, then compare performance in one dashboard. Keep execution consistent while diversifying venue risk.
Start by letting the model draft a plan, then convert it into measurable conditions inside Coinrule. Example: if price drops 5% within 4 hours, buy 0.5% of balance, repeat up to 3 times, then stop buying for 24 hours. Add an exit rule that sells 25% at +3%, 25% at +6% and trails the rest by 1.2%. You get creativity with the discipline of rule-based execution.
Rather than increasing size after one good day, validate the rule across multiple market phases. Check behavior during high volatility, low volatility and fast reversals. Scale position sizing gradually and keep a max exposure cap per asset.
When you want on-chain execution, use smart accounts with limited-permission sessions and defined spend limits. Routing can seek best liquidity across pools, while keys and approvals remain under your control. Keep the same risk framework you use on centralized venues.
Explore ToolsFAQ
Lars H.
Priya K.
Michael R.
Giulia C.
Kenji M.
Valentina C.
Every automated system needs assumptions written down. Define what invalidates the setup, such as a break below weekly support or funding staying extreme. Encode that as a hard exit or a pause condition.
Constraints make strategies more robust. Limit entries to a maximum of 2 per day, require at least 1% distance to stop, and block trades during major news windows you define. Small rules often cut noise more than adding indicators.
Let research generate signals, then let Coinrule handle execution. This keeps the process modular: swap an entry trigger without touching exits, sizing or risk limits. Debugging becomes much faster.
Track win rate, average win-to-loss ratio, maximum drawdown and time in market. Compare performance by exchange and by regime, not just total PnL. Use insights to adjust thresholds in small steps.
Once a rule set behaves as expected, scale slowly. Increase size in 10-20% steps, keep exposure caps, and review logs weekly. Consistency comes from process, not prediction.
Pick a template, connect an exchange with limited API permissions and start with small sizing. Add three essentials: a clear entry, a clear exit and a clear stop. Iterate from real data, not hunches.
Start Bot
