executeSmart()

Auto-routed multi-source execution — splits across protocols for maximum capacity and reliability.

What is executeSmart?

When you call execute() or executeLocal(), VAEA picks one protocol. But what if Marginfi has 50K SOL and Kamino has 80K SOL, and you need 100K? executeSmart() automatically evaluates capacity across all sources and routes your request optimally.

typescript

// Standard — fails if single source < 100K
await flash.execute({ token: 'SOL', amount: 100_000 }); // ❌ INSUFFICIENT_LIQUIDITY

// Smart — auto-splits across protocols
await flash.executeSmart({ token: 'SOL', amount: 100_000 }); // ✅ splits 50K Marginfi + 50K Kamino

How It Works

text

executeSmart({ token: 'SOL', amount: 100_000 })
  ↓
1. Query /v1/capacity → get available amounts per source
   → Marginfi: 50,000 SOL
   → Kamino:   80,000 SOL
   → Jupiter Lend: 35,000 SOL
  ↓
2. Route optimizer:
   → Need: 100,000 SOL
   → Plan: 50K from Marginfi (cheapest) + 50K from Kamino
  ↓
3. Build nested sandwich:
   begin_flash(SOL, 50K, marginfi)
   begin_flash(SOL, 50K, kamino)
   [your instructions]
   end_flash(SOL, kamino)
   end_flash(SOL, marginfi)
  ↓
4. Single atomic TX — all or nothing

Code Examples

const sig = await flash.executeSmart({
  token: 'SOL',
  amount: 100_000,
  onFunds: async (ixs) => {
    ixs.push(myLargeArbInstruction);
    return ixs;
  },
}, {
  maxSources: 3,              // max protocols to split across
  priorityMicroLamports: 5000,
});

console.log('Executed:', sig);

executeSmart() vs execute()

	execute()	executeSmart()
Sources	1 protocol	1-3 protocols
Max capacity	Single pool limit	Aggregate across all pools
TX size	Smaller	Larger (nested sandwich)
CU cost	~25K	~25K per source
Use case	Standard borrows	Large borrows, max capacity

💡 Tip

Use executeSmart() when your borrow amount might exceed the capacity of a single lending protocol. For standard amounts, execute() is simpler and uses less CU.