DexLadder Open the app

Build a bot, or bring your own

Two ways in: assemble a strategy from rules with no code, or plug in the bot you already wrote — as a sandboxed JavaScript file, or as your own program in any language answering on localhost.

Last verified 13 September 2026 · written by the DexLadder team

Build your own

A rules builder that composes entries, exits and risk constraints without code, running against live market data with paper capital. The point is not to avoid programming; it is to make the strategy explicit enough to be tested, because a rule you cannot write down is a rule you cannot measure.

Bring your own — two transports

TransportHow it worksWhy you would use it
Sandboxed JavaScriptYour strategy file is the Web Worker's script — so there is no eval anywhere and your code cannot touch the pageFastest path if the logic is already JS
Local signal endpointYour bot, in any language, serves BUY/SELL JSON at a localhost URL that DexLadder polls once per closed barKeeps your Python, Rust or Go bot exactly where it is

Isolated paper wallets

Every bot trades its own paper wallet with a stake you set. Your main paper account is untouched, so a bot that goes wrong cannot contaminate the record you are actually keeping. A bot that earns its place can be promoted later.

Ranked against the only benchmark that matters

The Arena ranks bots by edge over buy-and-hold, not by raw return. A bot that made 40% in a quarter where the asset made 60% did not find an edge; it found beta and charged you fees for it. Results carry the full risk picture — Sharpe and Sortino, the underwater curve, realised slippage — rather than a single flattering number.

Built to be stopped

A kill switch halts a bot through a planned unwind rather than an abrupt abandon, and every action lands in an execution logbook with endpoint telemetry, so a misbehaving signal source is visible rather than mysterious. Backtests run in their own worker, so the interface stays responsive while a strategy grinds through history.

Then test it properly

A bot that looks good on live paper should still be run through bar-by-bar historical replay, and judged with the cost stack from how fees actually work applied. Most bot strategies that fail in production fail on execution cost, not on logic.

More of the desk

Calculators for this

Terms used here

Questions people actually ask

Can I run my own trading bot against live data without an exchange account?

Yes. A bot connects either as a sandboxed JavaScript strategy or as a local endpoint DexLadder polls, and it trades an isolated paper wallet on live market data.

Is my bot code uploaded anywhere?

No. A JavaScript strategy runs as the script of a Web Worker in your own browser, and a signal endpoint runs on your own machine.

How are bots ranked?

By edge over buy-and-hold, with Sharpe, Sortino, the underwater curve and realised slippage shown alongside — not by raw return.

Open the desk — free, no account — no account, nothing to install