LaoLong

Crypto trading bots

How crypto trading bots work and how to choose one: a checklist

How crypto trading bots work, the main bot types, and a 7-point checklist to judge any bot: track record, custody, fees, drawdown, leverage.

A crypto trading bot is software that connects to an exchange and places trades automatically according to a strategy, such as buying dips on a grid, buying a fixed amount every week, or following trends. To choose one, judge it on evidence rather than marketing: a complete and verifiable trade history, clear custody of funds, all-in costs, and how deep its losses have been.

How a trading bot works

Under the hood, almost every bot has the same five parts:

  1. Connection. An API key or an agent wallet lets the bot place orders on an exchange account. The permissions on that key, whether it can only trade or can also withdraw, are the single most important security setting.
  2. Market data. Prices and volume, plus for futures the funding rate and open interest.
  3. Strategy. The rule that turns data into "buy", "sell" or "wait".
  4. Risk management. Position size, stop-loss, a maximum total exposure, and what to do in a crash.
  5. Execution and records. Orders are sent, fills and fees are logged, and profit or loss is calculated.

Bots don't get tired or emotional, and they can watch dozens of markets at once. But a bot only automates its strategy. If the strategy has no edge after costs, automation just loses money faster.

The main types of crypto trading bots

Bot type What it does Works best when Typical failure
Grid bot Places buy and sell orders at fixed steps in a price range Price moves sideways inside the range Price leaves the range; you hold a falling coin or miss a rally
DCA bot Buys a fixed amount on a schedule or on dips Long-term accumulation of assets you want to hold No exit logic; buys all the way down in a bear market
Trend / momentum bot Goes long in uptrends, short in downtrends Strong, lasting moves Choppy markets cause many small losses
Arbitrage / funding bot Captures price or funding differences between markets Spreads exceed fees Edge is thin; execution and counterparty risk
Signal / copy bot Executes trades from a signal provider or another trader The provider has a real, verified edge Provider's record is unverified or cherry-picked
AI bot Uses machine learning or a language model somewhere in the loop AI handles a task it is proven at, such as reading news for risk "AI" is a label, or the model overfits past data

Research supports trend and momentum more than most other approaches: Liu and Tsyvinski report "a strong time-series momentum effect" in crypto (Review of Financial Studies, 2021), and a 2025 study of Donchian-channel trend models on the 20 most liquid coins reports a net-of-fees Sharpe ratio above 1.5 in backtests (Zarattini, Pagani & Barbon, SFI Research Paper 25-80). A backtest is a starting point, not proof of live results. For more on AI specifically, see What is an AI crypto trading bot?.

The checklist: 7 things to verify before you deposit

1. A verifiable track record

Ask for every closed trade, not a highlights reel. A useful track record has:

  • a start date and an "as of" date;
  • the denominator ("41 of 50 trades closed in profit", not "82% win rate");
  • losing trades shown next to winning ones;
  • a written definition of what counts as a win;
  • a way to check entries independently: timestamped posts, on-chain transactions, or an exchange account you can inspect.

A backtest, a demo account or a screenshot is not a track record.

2. Custody: who controls the money

There are three common models:

Model How it works What to check
Your own exchange account + API key Bot trades on your account; you keep the money The key has trading only, no withdrawal permission; IP allow-list if available
Your own wallet on a decentralized exchange Bot signs trades via an approved agent key The agent key cannot withdraw; you can revoke it
Pooled or shared account Clients deposit into one account the service trades; each holds a share Who signs withdrawals, where funds sit, how shares are booked and reconciled, withdrawal rules

None of these is risk-free. What matters is that the provider tells you exactly which model it uses and what protects the funds.

3. All-in costs

Fees compound against you on every trade. On Hyperliquid, for example, the base perpetual futures fee is 0.045% for taker orders and 0.015% for maker orders (Hyperliquid docs: fees). A round trip with two taker fills costs 0.09% of position size. A bot that turns over its full position ten times a day pays about 0.9% of that position in fees every day, before it has earned anything.

Add funding payments on perpetual futures, spreads and slippage, plus any subscription or profit share charged by the bot itself.

4. Drawdown, not just returns

Maximum drawdown is the largest peak-to-trough fall in account value. It tells you what holding the strategy actually feels like. A strategy with a 40% maximum drawdown needs a 67% gain just to get back to even. Ask for the worst drawdown and how long recovery took.

5. Win rate in context

A high win rate means little on its own. If a strategy wins 90% of trades at +1% and loses 10% at −12%, it loses money: 0.9 × 1 − 0.1 × 12 = −0.3% per trade. Always look at the average win, average loss, and the sum of all results.

6. Leverage and stops

Does every position have a stop-loss that sits on the exchange, so it works even if the bot goes offline? What is the maximum leverage? Isolated or cross margin? Our guide to leverage, liquidation and position sizing shows how fast leverage shrinks the room for error.

7. Who is behind it, and what do they promise

The CFTC's advice is to "research the background of the company or trader" and to be wary of anyone who promises certain or unusually high returns; it cites a bot that "guaranteed at least a 10 percent monthly return" and was a Ponzi scheme (CFTC advisory). Crypto investment fraud cost Americans $5.8 billion in 2024 alone, according to the FBI (FBI IC3 2024 report, via Cointelegraph). Guaranteed daily or monthly returns are the clearest red flag there is.

Quick red-flag list

  • "Guaranteed", "risk-free", "X% per day".
  • Results only as screenshots, or only winning trades.
  • You are asked to send funds to a personal wallet or to recruit friends to unlock withdrawals.
  • No explanation of what the bot or its "AI" actually decides.
  • Withdrawals are delayed, or come with surprise fees.

How LaoLong AI Trading Bot measures up

Here is how LaoLong AI Trading Bot answers the checklist, in plain terms:

Checklist item LaoLong AI Trading Bot
What it trades Crypto perpetual futures on Hyperliquid, long and short, on liquid coins
Strategy A rules-based trading engine that follows trends; an AI risk officer moves the account between Normal, Caution and Defense
Custody model Shared trading account on Hyperliquid; your balance is your share, so every client gets the same percentage result
Key permissions The trading key can place orders but cannot withdraw; keys are kept in an isolated, encrypted signing service
Withdrawals Free balance can be withdrawn at any time to your own address; money in open trades unlocks when they close; withdrawals go only to allow-listed destinations under an hourly cap
Stops Trades on the AI's calls use isolated margin with a stop order resting on Hyperliquid; the trend book has an emergency stop checked every minute
Track record Every closed AI trade call is published, losses included: 214 calls closed in profit; win rate of the public calls 78.1%, as of 17 Nov 2026, 15:34 UTC. Each public call links to its post
Costs Exchange fees and funding are already included in every trade result we report
Minimum deposit $10

That track record measures our AI trade calls as price moves without leverage. The methodology page defines every number, and the results page lists every closed call.

Bottom line

Choose a trading bot the way you would choose a fund manager: check the full record, the custody, the costs and the worst drawdown, and walk away from anyone who promises a fixed return.

Questions and answers

How does a crypto trading bot work?

It connects to an exchange through an API key, reads market data, applies a strategy such as grid, DCA or trend following, and places orders automatically. A risk layer sets position size and stops.

Are crypto trading bots profitable?

Some are, many are not. A bot only automates a strategy; if the strategy has no edge after fees, automation loses money faster. Judge a bot by its complete, dated trade history after fees, including losses and drawdowns.

What is the most important thing to check before using a bot?

Custody: who controls the money and whether the key the bot trades with can withdraw it. Second, a verifiable track record that includes every losing trade.

Is a high win rate a good sign?

Only together with the size of wins and losses. A strategy that wins 90% of the time can still lose money if the average loss is ten times the average win.