Skip to main content
You shouldn’t need to understand smart contract parameters to execute a DeFi strategy. SURCHI’s Natural Language Interface (NLI) lets you describe what you want to do in plain English — and translates your intent into a fully structured, executable strategy. No manual parameter entry, no protocol-specific syntax. Just say what you want, review what the system understood, and approve it to activate.

​
How Natural Language Commands Work

The NLI is built around an interpretation-first, execution-second model designed to keep you in control at every step:
  1. You describe your intent in plain English — as specific or as general as you like.
  2. The NLI parses your command and constructs a structured strategy object: asset targets, amounts, trigger conditions, timing, DEX routing preferences, and slippage limits.
  3. The interpreted strategy is presented to you in full — every parameter rendered clearly before anything is queued or executed.
  4. You confirm, amend, or discard the generated strategy. Nothing executes without your explicit approval.
The NLI is connected to all three Sentinels. A trading command routes through Alpha → Liquidity → Execution. A monitoring command activates Alpha Sentinel watchlists. A liquidity protection command configures Liquidity Sentinel thresholds. The system determines the right path from your intent.

​
Example Commands

​
Trading & Swaps

Buy $200 worth of SOL if it drops below $90

Sell half my BONK when it reaches a 40% gain from my entry

DCA into JTO with $50 every Monday at 9am UTC

​
Liquidity & Yield

Provide $500 liquidity to the SOL/USDC pool on Orca if APY is above 20%

Withdraw my liquidity from any pool if impermanent loss exceeds 10%

​
Monitoring & Alerts

Alert me if any wallet holding more than 1 million BONK sells more than 20%

Notify me when SOL funding rate turns negative

Watch the SURCHI/SOL pool for any TVL drop over 15% in one hour

​
Strategy Confirmation Flow

Every natural language command follows the same structured confirmation process before anything happens on-chain:
1

Enter Your Command

Type your intent in natural language into the SURCHI NLI input — from the dashboard, mobile interface, or API. No special syntax required.
2

SURCHI Parses and Displays the Strategy

The NLI interprets your command and generates a full strategy object. Every parameter is displayed clearly: asset targets, trigger conditions, amounts, timing, DEX routing, slippage tolerance, and expiry. You see exactly what the system understood before making any decision.
3

Review and Amend

Inspect the generated strategy. If any parameter isn’t what you intended — the amount, the trigger threshold, the pool — you can edit individual fields directly without re-entering the full command. Adjust and re-validate as needed.
4

Approve to Activate

Confirm the strategy. It is submitted to the Sentinel pipeline: Liquidity Sentinel validates risk conditions, and Execution Sentinel monitors for trigger conditions and acts when they are met. The strategy remains active until triggered, expired, or manually cancelled.

​
Command Tips

Getting the most out of the NLI is about precision. The more specific your command, the more accurately the system can translate your intent:
  • Be specific about amounts — distinguish between dollar amounts (“$200 worth of SOL”) and token amounts (“200 SOL”). Ambiguity defaults to the dollar interpretation.
  • Include time conditions when relevant — “by end of week”, “within the next 4 hours”, or “expires Friday 5pm UTC” all help the system set appropriate strategy windows.
  • Specify which DEX or pool if you have a preference — “on Orca”, “via Raydium”, or “using the SURCHI/SOL pool” route your strategy to the correct venue. If unspecified, the system routes for best execution.
  • Set explicit stop-loss conditions for automated strategies — any strategy that involves autonomous execution should include a protective condition: “stop if price drops below X” or “cancel if slippage exceeds 1%”.
The NLI works best with specific, quantified commands. The more parameters you include — amounts, thresholds, timeframes, venues — the more precise the generated strategy, and the less you’ll need to amend at the confirmation step.