Bitsgap VS Trade Leopard

Trade Leopard's answer:

Because it does less, on purpose.

Most of this category converged on the same stack — grid bots, signal marketplaces, copy trading, arbitrage, futures. Trade Leopard is one allocation policy that you configure. If you want a marketplace of strategies to rent, this is the wrong product.

What you get instead:

• A rule you can state in a sentence. Target weights, a budget, a pace, and a condition each asset has to meet before its share is spent.
• Buying that corrects your allocation instead of selling to fix it. New money goes to whatever is furthest below target.
• Nothing sold behind your back. Out of the box the engine only buys. The single automatic sell is the optional trim, it ships off, it sells profit only, and it never touches a position at a loss.
• Orders you see before they happen. Lump sums, percentage sells and rebalances all preview the exact orders asset by asset — including any leg skipped for falling under the exchange's minimum — before anything is placed.
• Keys that cannot move your money. You create the API key without withdrawal permission, and every order we place leaves from one fixed IP address, so you can lock the key to it.

And one honest reason not to choose it: we make no performance claim. There are no backtests on our site, because the product makes no edge claim it would have to defend. If a comparison table of returns is what decides it for you, we will lose that comparison by not entering it.

Trade Leopard's answer:

Two pieces, with a database between them.

• Dashboard — Next.js (App Router) and React, deployed on Vercel. Charts are drawn with lightweight-charts. Billing is Stripe. Alerts go out as web push, and the dashboard installs to a phone home screen as a PWA.
• Engine — a Python worker that holds the strategy loop and places the orders. It is headless and exposes no public URL, so the process holding the exchange keys has no inbound attack surface.
• Shared state — Supabase for authentication, settings, encrypted secrets and the trade log. The website never trades directly: it writes a row telling the engine what to do, and the engine acts on it.
• Exchanges — the Kraken, OKX and Coinbase REST APIs, called with keys created without withdrawal permission.

The split is deliberate. The half a browser can reach cannot place an order, and the half that can place orders cannot be reached from the internet.

Trade Leopard's answer:

Three things, and they are all about restraint rather than features.

It buys on a condition, not a date. You set a budget and a pace, but each asset's share is only spent when that asset clears a gate you configure — RSI across four timeframes (1h, 2h, 4h, 1d), with you choosing how many have to be under your threshold. Anything that does not qualify is skipped, and the money it did not spend rolls into the next period instead of being pushed somewhere weaker. Over time that accumulates cash for the next real dip.

Rebalancing is buy-only. The standard way to correct a drifted portfolio is to sell whatever grew past its target. Trade Leopard sends new money to whatever sits furthest below target instead. Nothing is ever sold to correct drift on its own, so the position that was working is never cut.

The optional trim only ever sells profit. Switch it on and an overbought holding that is up by at least the percentage you set has its gain sold — never the money that bought it, and never anything at a loss. It ships switched off.

One more thing worth stating plainly: we publish no backtests and make no claim to outperform. It is an allocation policy you set, not a strategy that promises returns.

Trade Leopard's answer:

People who already know which coins they want to hold, and want to accumulate them on a rule instead of on a mood.

Typically that means someone who:

• Has a handful of assets they have researched and intend to hold for years, not a watchlist they rotate weekly.
• Adds money regularly and wants it split along weights they chose rather than dumped into whatever ran last.
• Keeps funds on their own exchange account and does not want a third party holding them.
• Is comfortable creating an API key and ticking the right permissions — the product walks through it for Kraken, OKX and Coinbase, but it assumes you are willing to do that step.

It is a poor fit for day traders, anyone looking for signals or strategies to copy, and anyone who wants leverage or futures. It has none of those and is not going to add them.

Trade Leopard's answer:

I've been in crypto since 2016. In that time I've used most of the platforms and most of the tools, traded on margin and with leverage, and worked just about every timeframe there is.

What I came away with wasn't a better indicator. It was that almost every bad decision I made started the same way — buying because something was moving and I felt like I was missing it, or selling because a headline got to me. A regulatory story, an exchange going under, another country supposedly banning it. News like that pushes people out of the market at exactly the wrong moment, and for years I was one of them. The setup didn't matter. The emotion was the constant.

So I stopped trying to time it and started deciding in advance. Pick the assets on their own merits. Decide the weights. Decide how much goes in and how often. I settled on how I wanted to approach the market before the market gave me a reason to panic, so I'm not shaken in the heat of the moment — the way a soldier is conditioned for the scenario long before he's in it.

Trade Leopard is that decision, running on my own exchange account. It buys my basket on the schedule I set, but only when an asset is actually oversold. It corrects my weights by buying rather than selling. Switch on the trim and it takes profit and leaves the stake invested. The exciting part is what it took away: the stress. Something that used to take up my whole day is now something I check when I feel like it.

I built it for myself first. The earlier version had a strategy I was fond of, an RSI "snowball" ladder. I retired it after three and a half years of backtesting put its profit factor well under 1. That's the standard I hold the product to: if I can't support it, I don't use it.