Blueprint DSL

See the Strategy Language in Action

You can inspect the blueprint language, but you do not have to write it by hand. SurgePilot AI agents understand it and can build validated strategies from plain-language trading ideas.

How It Works

From plain-language idea to deterministic bot behavior.

Tell your AI agent how you want the strategy to behave. It can write the SurgePilot blueprint, choose the right execution profile, and expose trader-facing settings. The runtime then validates the strategy and executes it deterministically.

AI-Agent Fluent

Your agent can translate plain-language strategy ideas into SurgePilot blueprints.

Profile Aware

Agents can target grid, candle-step, tick-reactive, and hybrid overlay execution.

Checked Before Live

Every blueprint is validated by the runtime before it becomes a bot strategy.

Deterministic Runtime

The bot executes checked strategy logic, not free-form AI output. SurgePilot's runtime engine removes AI execution costs that other systems incur.

From Idea to Bot
strategy → parameters → rules → diagnostics
parameters

Trader-facing controls with safe defaults, labels, ranges, and optimizer hints.

indicators

SMA, EMA, RSI, MACD, Bollinger Bands, ADX, ATR, momentum, volatility, highs, and lows.

state

Strategy memory for armed setups, profit peaks, cooldowns, and counters.

rules

Readable entry, exit, and update conditions for the behavior the bot should follow.

diagnostics

Live values shown in the product so traders can see what the bot is watching.

optimization

Marks which controls can be tuned safely in backtests or by the auto-optimizer.

Example Blueprints

Real strategy ideas your AI agent can generate.

Each example shows how a natural-language trading request can become a testable, explainable strategy for a supported SurgePilot execution profile.

Candle-Step Signal
Pullback and rebound logic

RSI Mean Reversion

Shows how an AI agent can turn a familiar pullback idea into a complete bot with controls, entries, exits, and live diagnostics.

strategy "RSI Mean Reversion" {
  description "Buys after price has fallen enough to look stretched, then exits after momentum recovers."
  family signal
  execution_profile candle_step

  parameters {
    param oversold_level number default 30 min 10 max 45 step 1 label "Buy When RSI Falls Below" description "Lower values wait for a deeper pullback before buying."
    param recovery_level number default 52 min 45 max 70 step 1 label "Sell When RSI Recovers Above" description "Higher values wait for a stronger rebound before selling."
    param position_size_pct number default 10 min 1 max 50 step 1 label "Cash To Use Per Buy %" description "How much available cash the strategy can use for one buy."
  }

  indicators {
    indicator rsi type rsi {
      period 14
    }
  }

  state {
  }

  rules {
    enter long when all {
      flat
      indicator_ready rsi.value
      rsi.value < oversold_level
    } size capital_pct position_size_pct reason "RSI shows a stretched pullback"

    exit long when all {
      long
      rsi.value > recovery_level
    } sell all reason "RSI recovered"
  }

  diagnostics {
    show current_rsi = rsi.value label "RSI"
  }
}
grid_ladder

Grid Ladder

An AI agent can turn a range-trading idea into planned levels, allocation, fills, and rebounds.

candle_step

Candle-Step

An AI agent can build candle-updated trend, pullback, DCA, and indicator-driven strategies.

tick_reactive

Tick-Reactive

An AI agent can build fast strategies that react to live pressure, spread, liquidity, and choppiness.

tick_reactive_hybrid

Hybrid Overlay

An AI agent can add tick-level confirmation to a base strategy before entries or exits execute.

Strategy library

See how these examples map to trader-facing strategy families.

Explore Strategies