You can build a Solana trading strategy by typing what you want in plain English — no code, no config files, no bot framework. For most of copy trading's history, encoding a rule like "follow this wallet, but skip anything with thin liquidity and sell when they sell" meant Pine Script, a scripting bot, or a settings file most people never open. Soltrace lets you describe that in a sentence and get back a real, editable strategy — then you check it before a single dollar moves.
This guide explains what "plain English" can actually express, what those descriptions turn into, and — the part that matters most — why the draft the AI hands you is a starting point, not a finished edge. If you want the fuller tour of the tool itself, see the Solana copy-trading strategy builder; this post is about the no-code way in.
Why plain English lowers the barrier
The reason most people never customise a copy-trading strategy isn't that they lack opinions about risk. It's that turning those opinions into rules used to require a second skill — coding — that has nothing to do with trading.
- Risk rules are ideas, not syntax. "Don't copy tokens younger than a day" is a clear thought. Forcing it through a scripting language adds a translation step where most people give up and just run the defaults.
- Defaults are someone else's risk tolerance. When the barrier to changing a rule is high, you inherit whatever the tool shipped with. Lowering the barrier means the strategy reflects your appetite, not a template's.
- Iteration gets cheap. If rewriting a rule is a sentence rather than a code change, you'll actually try three versions and compare them — which is how you find something that fits, instead of settling for the first thing that runs.
Plain English doesn't make you a better trader. It removes the reason a non-coder had to stay a passenger.
What plain English can actually express
A copy-trading strategy is a small set of decisions — who you follow, what you let through, how much you commit, and when you exit. Plain-English input can reach every one of them. In practice a good description touches:
- A target wallet. The address (or addresses) you want to copy. This is block one, and everything downstream adapts its behaviour to your rules.
- Entry care. The filters that decide which of the target's buys you actually take — skip low-liquidity pools, skip brand-new mints, require the mint and freeze authorities to be revoked, cap the price impact your own buy would cause. These are the rug-avoidance rails most people wish they'd set earlier.
- A sizing rule. How big each copy is. You can say it in words — a fixed amount of SOL, a fixed percentage of your vault, proportional to the target's trade, or proportional to your whole portfolio — and it maps to one of the four sizing modes.
- Exits. When to close: take profit at a multiple (in tiers if you want), a stop-loss, a trailing stop, scaling out in steps, closing after a set number of hours, or simply selling when the target sells.
If you can describe those four things in a sentence or two, you've described a strategy. The AI's job is to turn the description into rules you can then see and edit — not to guess at what you left out.
Three example prompts and what they become
These are illustrative and clearly hypothetical — they show the shape of the translation, not numbers you should copy. Your own settings depend on the wallet and the variance you can stomach.
-
"Copy this wallet, only tokens with revoked authorities and decent liquidity, 0.5 SOL a trade, take profit at 2x and stop out at 40% down." Becomes: one target wallet; entry filters for authorities-revoked and a minimum-liquidity floor; fixed-amount sizing at 0.5 SOL; a take-profit at 2x and a stop-loss at -40%.
-
"Follow these three wallets but skip anything under a day old, size each copy at 5% of my vault, trail the stop by 25%, and sell whenever they sell." Becomes: three targets; a minimum-token-age filter; fixed-percentage-of-vault sizing at 5%; a 25% trailing stop; and mirror-the-target's-sells layered on top so you're not left holding after they've exited.
-
"Mirror this trader's conviction, only low price-impact entries, bank half at 3x and let the rest ride, and close anything still open after 12 hours." Becomes: proportional-to-target sizing so their big swings become your big swings; a max-price-impact filter; a tiered take-profit (scale out half at 3x, trail the remainder); and a 12-hour time-based exit.
Notice what each prompt did: it picked a wallet, drew an entry boundary, chose one sizing mode, and stacked a couple of exits. That's the whole grammar. The English is just a friendlier surface over the same building blocks every strategy is made of.
The honest part: a draft, not an edge
Here's the thing the "AI builds your strategy" pitch usually leaves out. The AI gives you a starting draft, not a finished edge. It's genuinely good at translation — turning your sentence into the correct filters, the right sizing mode, sane exit rules. It has no idea whether those rules would have made money on the wallet you chose.
So the plain-English step is the beginning of the work, not the end of it. Once you have a draft, you do two things:
- Review it in Simple or Advanced. Read back what the AI built and adjust anything that doesn't match your intent. A description is lossy; the editable strategy is where you make it exact.
- Backtest it before you fund. Replay the whole thing against the target wallet's real on-chain trades, priced on 1-second data, and see the equity curve and the max drawdown it would have produced. That's the difference between a strategy you believe in and one you've checked — the backtesting mechanics are here.
Skipping straight from a nice-sounding prompt to a funded vault is exactly the mistake the tool is built to prevent. The English gets you a draft fast so you have more time for the review and the test — not less reason to do them.
Simple vs Advanced: two ways to tune the same strategy
The draft opens into one of two views, and you can move between them.
- Simple is presets and guardrails. Sensible defaults, a smaller set of dials, and rails that keep a setting from doing something obviously dangerous. It's the right view when the AI's draft is close and you want to nudge a couple of numbers without touching everything.
- Advanced exposes every knob. Each filter threshold, the full sizing configuration, every exit rule and its tiers — all editable by hand. It's for when you want precise control over a rule the plain-English pass approximated, or when you're comparing tightly tuned variants.
Both views describe the same underlying strategy, so nothing is locked behind the mode you start in. You can begin in plain English, refine in Simple, and drop into Advanced for the one rule you care about — it's all the same strategy, and it stays yours throughout.
From a sentence to a live vault
Once the draft reads the way you want and the backtest holds up, going live is short. The strategy you described — the target, the entry care, the one sizing mode, the exits — is what the keeper runs forward. Your funds stay in a non-custodial Solana PDA vault you own the entire time: the keeper can place and size copies but can never withdraw. Execution is keeper-signed across the Jito, Nozomi, Astralane and AllenHark relays, with optimistic execution on Pro and Elite, at a flat 0.85% per trade and no profit share. You typed the idea; the vault, the funds, and the final say stay with you.
Ready to describe a strategy instead of coding one? Open the strategy designer, say what you want in plain English, then edit and backtest it before you fund a thing.
Copy trading does not guarantee profit, and crypto trading carries substantial risk of loss. An AI-generated draft is a starting point, not a prediction; nothing here is financial advice.