Automate Crypto Trading Without Code or Servers

TL;DR: Automating a crypto strategy used to mean writing a script, renting a server, wiring an exchange API and hoping nothing crashed overnight. You can now do it without code or infrastructure: define the rules visually or in plain English, backtest them with realistic assumptions, forward-test on a paper agent, then run a live agent that trades on Hyperliquid through a key that cannot withdraw. On Trigr the whole path lives in one place, and the agent keeps running while your browser is closed.

What does "automating a strategy" actually involve?

A trading bot is more than a buy rule. To run one reliably you need all of these pieces:

1. Rules that turn market data into entries and exits.
2. Data for every input, available at the moment the rule is evaluated.
3. A backtest that shows how the rules would have behaved, with realistic fills and costs.
4. Execution that places, protects and closes orders on an exchange.
5. Hosting so the bot runs around the clock, restarts after failures and survives your laptop sleeping.
6. Key security so a compromised bot cannot empty your account.
7. Monitoring so you learn about problems before they become losses.

Code-first setups make you own all seven. Many "no-code" bots cover only rules and execution, with little or no honest backtesting. The useful question when choosing an approach is which of these you want to own yourself.

Step 1: How do you turn an idea into rules without code?

Most trading ideas are already rules in disguise. "Buy ETH when it breaks out on the 4-hour chart, but only if Bitcoin is in an uptrend, and get out if it drops 3%" contains a trigger, a filter, a direction and a risk rule.

Trigr's Studio represents exactly that structure as a node graph: one TRIGGER, optional FILTERs, a SIGNAL that sets direction, a RISK node for sizing and exits, and EXECUTE. Any node can read a different market than the one you trade, so the "only if Bitcoin is in an uptrend" part is one filter node pointed at BTC. The no-code strategy builder guide walks through each node.

You do not have to place nodes by hand. You can describe the idea in plain English to the Studio copilot and let it draft or edit the graph, then check what it built. From plain English to a backtested strategy covers what to verify before trusting an AI draft. If you prefer your own assistant, ChatGPT, Claude or Codex can connect to Trigr over MCP and build strategies through the same validators.

The RISK node is where automation gets its discipline. It sets position size (1 to 100% of the slot's capital), leverage (1 to 50x), maximum concurrent positions, take-profit and stop-loss, direction, an ATR trailing stop, a time stop and whether to exit when the signal flips. Pick conservative values first; the position sizing guide explains how they interact.

Step 2: How do you know the rules are worth automating?

Automation executes a bad strategy just as faithfully as a good one, only faster. A backtest is the first filter, and only if it is honest.

Trigr's engine is built to avoid the usual ways backtests flatter a strategy:

• Point-in-time data. Every input uses only information available at the bar's close.
• Next-bar-open fills. A signal fills at the next bar's open, not at the price that generated it.
• Conservative intrabar handling. If a take-profit and a stop-loss are both touched inside one bar, the engine replays 5-minute bars to find which came first, and if it is still ambiguous, the stop wins.
• Explicit costs. Trading fees are always applied. Slippage and funding are opt-in and off by default, so a first result is labeled gross; switch them on before deciding anything.

A standard backtest costs 50 credits, and the Free plan includes 3,000 one-time trial credits. Read the trade log, not just the equity curve, and be suspicious of any result that only works on a short stretch of history.

Step 3: Why paper trade before going live?

A backtest tells you what would have happened. A paper agent tells you what does happen now, with live prices and the real signal timing. It runs the same pipeline as a live agent but fills in a virtual portfolio at the current Hyperliquid price, deducting the base taker fee and builder fee. It does not model slippage or funding, so treat it as a check on behavior rather than a forecast of net returns.

Paper agents need nothing from you beyond capital figures and strategy choices: no exchange connection, no server. An agent can hold up to six strategies, each trading a different asset with its own allocation, and the setup screen shows a combined backtest and a risk panel before you start. Paper trading agents explains how long to run one and what to compare.

Step 4: How does a live agent run without your computer?

This is where no-code platforms differ most, and where the details matter.

On Trigr, a live agent trades on Hyperliquid or on a Propr prop-firm challenge account. For Hyperliquid, you approve a separate trade-only key with your wallet. Hyperliquid's API wallet model lets such a key trade on your account while withdrawals still require your master wallet's signature.

For an autonomous agent to keep trading while your browser is closed, Trigr stores that trade-only key encrypted on its side. It can place and cancel orders, but it cannot withdraw or send funds to another address. You keep your master wallet and custody of your funds. The key is valid for about 90 days, and you can revoke it from your master wallet at any time.

Once live, the agent handles the operational work a script would otherwise need:

• Each entry ships with reduce-only take-profit and stop-loss orders on the exchange, retried and repaired if placement fails.
• A reconciliation job checks the agent against Hyperliquid about every minute and attributes each close from real fill data.
• Pausing blocks new entries while existing positions can still close.

Two practical constraints to know upfront: Hyperliquid rejects orders under $10 of notional, so Trigr raises smaller agent entries to $11, and an agent with less than $11 of capital cannot open positions. And TradFi markets on Hyperliquid (such as gold or the S&P 500) are backtest and paper-only in the current beta. The deployment guide covers funding, subaccounts and sizing in detail.

Step 5: How do you keep an eye on it?

No-code does not mean no attention. Turn on alerts for fills, take-profits and stop-losses, and agent errors, by email or web push. Compare the live record against a backtest run with costs switched on. Decide in advance what would make you pause, such as a drawdown beyond the backtest's worst or a trade frequency far from what the backtest showed.

What do you give up by not writing code?

A no-code platform is a trade-off, and it is worth being clear about it:

• You work within the available building blocks. Trigr supports a broad set of indicators and data sources, but not arbitrary custom logic. Exits come from the RISK node options; there are no custom exit graphs.
• You work within the supported markets and venues. That means 43 certified crypto perps on Hyperliquid plus HIP-3 TradFi markets for research and paper trading, with live execution on Hyperliquid or Propr.
• You work on bar timeframes. Strategies run on 5-minute to monthly bars, not tick data, which rules out high-frequency and market-making strategies.

If your idea needs custom order-book logic, tick-level execution or a venue Trigr does not support, writing code is the right choice. Coding your own backtester vs using Trigr discusses honestly when code wins.

What does it cost?

There are no spreads or markups on orders. Per trade you pay Hyperliquid's own fees plus a Trigr builder fee of up to 0.05%, collected by Hyperliquid through its builder code mechanism after you approve it once. Live trading is available on every plan, including Free; paid plans add monthly credits, more saved strategies and more paper agents. Current plans are on the pricing page.

Automation removes manual effort, not risk. Backtests are not guarantees, and leveraged perpetual futures can lose more than you expect.

Next steps

Start with the getting started docs, build one simple strategy, and run it on a paper agent before connecting any funds.