PATAS

Pattern Analysis for Text Anti-Spam

Endpoint

POST /v1/patas/analyze

Request Body

{
  "messages": [
    {"content": "Free V-Bucks!", "label": "spam"},
    {"content": "Hello world", "label": "ham"},
    // ... more messages
  ],
  "options": {
    "min_confidence": 0.8,
    "mode": "SAFE_AUTO" // "SAFE_AUTO" | "REVIEW_ONLY"
  }
}

Response

{
  "success": true,
  "rules": [
    {
      "action": "BLOCK",
      "pattern": "free.*v-?bucks",
      "confidence": 0.94,
      "cluster": "gaming_scam",
      "examples_matched": 23
    }
  ],
  "mode": "SAFE_AUTO",
  "credits_used": 5
}

Modes

SAFE_AUTO

High-confidence rules auto-block. Recommended for production.

REVIEW_ONLY

All matches flagged for human review. Good for testing.

Pricing

$0.10 per 1,000 messages analyzed