📨 Single Message Test (GUVI Format)

Test the exact GUVI input format for a single scam message

📝 Request Builder

Quick Examples:

📋 GUVI Format Request

{
  "sessionId": "test-session-001",
  "message": {
    "sender": "scammer",
    "text": "Your message here...",
    "timestamp": 1738756800000
  },
  "conversationHistory": [],
  "metadata": {
    "channel": "SMS",
    "language": "English",
    "locale": "IN"
  }
}

📬 API Response

📭

Send a request to see the response

💬 Multi-Turn Conversation Test

Test multi-turn conversations with conversationHistory

🗨️ Conversation

Turn: 0
💬

Type as Scammer → Your AI responds as Victim

This simulates how GUVI will test your honeypot

🦹 Scammer:

📊 Session State

-
-
-
0
-
-

🎯 Extraction Progress

0%
❌ UPI ❌ Bank ❌ Phone ❌ IFSC
⏳ Waiting for extraction...
Callback triggers when confidence ≥ 85% or turn 20

📦 Extracted Intelligence

No intelligence extracted yet

🔑 Agent Notes

-

📋 conversationHistory (sent to API)

[]

📤 GUVI Callback Test

Test the mandatory final result callback format

📝 Callback Payload Builder

📋 GUVI Callback Format

{
  "sessionId": "test-session-001",
  "scamDetected": true,
  "totalMessagesExchanged": 18,
  "extractedIntelligence": {
    "bankAccounts": [],
    "upiIds": [],
    "phishingLinks": [],
    "phoneNumbers": [],
    "suspiciousKeywords": ["urgent", "verify now", "account blocked"]
  },
  "agentNotes": "Scammer used urgency tactics..."
}
POST https://hackathon.guvi.in/api/updateHoneyPotFinalResult

⚠️ Callback Rules

🔑

When to Send

  • Scam intent is confirmed (scamDetected = true)
  • AI Agent has completed engagement
  • Intelligence extraction is finished
⚡

Trigger Conditions

  • Max turns (20) reached
  • High extraction confidence (≥85%)
  • Session explicitly terminated
🚨

MANDATORY

  • This callback is required for scoring
  • Without it, solution cannot be evaluated
  • Platform measures engagement depth

✅ GUVI Compliance Checklist

Verify your implementation meets all GUVI requirements

🔌 API Requirements

⏳

x-api-key Authentication

API accepts and validates x-api-key header

⏳

GUVI Input Format

Accepts sessionId, message object, conversationHistory, metadata

⏳

GUVI Output Format

Returns status and reply fields

⏳

Response Time < 2s

API response time within acceptable limit

🎯 Detection Requirements

⏳

Scam Detection

Detects scam intent with confidence score

⏳

Multi-Language Support

Handles English, Hindi, Hinglish

🤖 Agent Requirements

⏳

Multi-Turn Conversations

Handles conversation history correctly

⏳

Human-like Responses

Agent behaves like real human

Manual Check

📊 Intelligence Extraction

⏳

UPI ID Extraction

Extracts UPI IDs like scammer@paytm

⏳

Bank Account Extraction

Extracts bank account numbers

⏳

Phone Number Extraction

Extracts phone numbers

⏳

Phishing Link Detection

Identifies suspicious URLs

📤 Callback Requirements

⏳

Callback Payload Format

Uses correct camelCase field names

Code Review
⏳

Callback Trigger Logic

Sends callback at appropriate times

Code Review