Your agent writes
the strategy.

State a stance in plain words. SuperChart researches, writes the rule, optimizes it against history, and runs it capped on-chain.

Launch appBack to overview
Input
Plain words
Output
Optimized rule
Rule
Yours to edit
Optimizing…
✦ Baseline RSI dip-buy
+41% return-23% max DD1.1 Sharpe284 trades
ResearchLive data, one report
DescribeWords become a rule
ReadRSI · EMA · trend
BacktestRun against history
CapOn-chain, revocable
01Deep Research

The agent does
the research.

Deep Research Mode gathers real price, technicals, on-chain data, news, and sentiment, then synthesizes one Trading Report. Open a section to read the working behind the call.

02Describe

Plain words become
a typed rule.

You speak in plain words. The agent turns intent into a typed strategy you can read, edit, and backtest. No indicator math by hand.

You, in plain words

“Buy SUI when it is oversold but the trend is up, with a stop, and size it sensibly.”

strategy.sup
if rsi(14) < 30 and ema(9) > ema(21):
    long(stop = atr(14)*1.5, size = kelly())
# capped 500 USDC/day, revoke anytime
03Read the chart

It reads RSI, EMA,
and the trend.

The agent pulls indicators straight off the price series and grounds every rule in what the chart is doing right now.

  • priceWhere price actually isThe same opens, closes, and wicks the chart draws are what the agent reads.
  • rsiWhen a move is stretchedRSI under 30 reads oversold — the rule above waits for the snap-back instead of chasing.
  • emaWhich way the trend leansFast over slow EMA frames whether a dip is a discount or a falling knife.
indicator readout · demo
RSI 1428 / oversold
03070100
EMA 9 vs 21+1.4% / above
belowflatabove
signal → buy candidate
04The dry run

Backtested before
it goes live.

Every rule runs against past candles before a coin moves. This is the optimized rule from the studio above. Fees modelled, slippage not.

BacktestSUI / USDC · 4h · rsi + ema + atr + kelly · illustrative
Total return+148%
Max drawdown-9%
Sharpe2.2
Total trades171
Demo valuesRule-based backtest on historical candles. Past performance does not predict the future. Research, not financial advice.
05Capped on-chain

It never trades
past the allowance.

A deployed strategy runs inside the same typed, capped grant as everything else in the wallet. Execution stops at your limit, and you revoke in one tap.

  • capBounded by budgetDaily, per-trade, and per-coin caps. The strategy never spends a unit past them.
  • atomicSettles in one blockA multi-step trade commits together. If one leg fails, nothing settles.
  • revokeOff in one tapPull the grant and the agent stops at the next check. The vault answers only to you.
St
Mean-revert strategy
grant 0x9f3c..a12e
Active
Daily spend120 / 500 USDC
Per-trade cap50 / 100 USDC
CetusDeepBookwindow: 30d
Enforced on-chain

Trade in plain words.

Describe the stance; the agent turns it into a tested, capped rule. Revoking takes one tap.

Launch appBack to overview