Live Playground

Execute skills in mock, real, or hybrid mode with structured outputs.

The Live Playground executes a selected skill against a protocol adapter.

It supports testing, debugging, and live data validation.

Execution modes

mock

Returns simulated output with realistic response structure.

This mode avoids live network calls.

real

Uses live protocol calls and mainnet data.

This mode can return real execution output or live protocol responses.

hybrid

Uses live market or network data with simulated execution.

This is useful for safe testing with current conditions.

Execution request shape

{
  "skillSlug": "jupiter-swap",
  "skillName": "Jupiter Swap",
  "protocol": "jupiter",
  "action": "swap",
  "mode": "mock",
  "input": {
    "amount": 1.0,
    "wallet": "7xKX...abc"
  }
}

Execution response shape

{
  "success": true,
  "mode": "mock",
  "output": {
    "success": true,
    "protocol": "Jupiter",
    "action": "swap",
    "result": {
      "source": "LIVE Jupiter Quote API v6",
      "inputMint": "So111...112",
      "outputMint": "EPjF...t1v",
      "outAmount": 89.14,
      "priceImpact": 0.001,
      "route": "Jupiter (Direct)"
    },
    "logs": ["[engine] Mode: SIMULATED"],
    "latencyMs": 312,
    "timestamp": "2026-03-13T08:41:00.000Z"
  },
  "executionId": 4201
}

Returned execution details

Every run can include:

• Success flag

• Mode used

• Structured protocol result

• Execution logs

• Measured latency

• Timestamp

• Transaction signature when available

Typical use cases

• Validate data shape before integration

• Test live quote behavior

• Inspect adapter logs

• Compare mock and real outputs