API Documentation

Integrate risk detection into any application. Send customer text, get back structured risk assessments in under 2 seconds.

Overview

The RiskDetect API analyzes customer interactions (emails, call transcripts, chat messages) and returns a risk assessment with a numerical score, detected signals, summary, recommended action, and urgency level.

All responses are JSON. All endpoints accept POST requests with JSON bodies.

Authentication

API keys are included in the Authorization header:

Authorization: Bearer rd_your_api_key_here

Get your API key from the dashboard settings. The live demo does not require authentication.

Base URL

https://riskdetect.app/api

Analyze Interaction

The primary endpoint. Accepts any customer text and returns a full risk assessment.

POST /api/analyze

{
  "content": "This is my third email about this issue...",
  "type": "email",           // optional: "email", "call_transcript", "chat"
  "subject": "RE: Urgent",   // optional
  "from": "customer@co.com", // optional
  "customerName": "Sarah J"  // optional
}

Request Parameters

Parameter	Type		Description
content	string	required	The customer interaction text to analyze
type	string	optional	Interaction type: email, call_transcript, chat. Auto-detected if omitted.
subject	string	optional	Email subject line (provides additional context)
from	string	optional	Sender email or identifier
customerName	string	optional	Customer name for personalized recommendations

Response

{
  "analysis": {
    "risk_score": 85,
    "risk_level": "high",
    "signals": [
      "BBB complaint threatened",
      "Legal counsel mentioned",
      "Competitor comparison"
    ],
    "summary": "Customer is highly frustrated...",
    "recommended_action": "Escalate to senior manager...",
    "urgency": "today",
    "category": "email"
  },
  "metadata": {
    "model": "gpt-4o-mini",
    "processingMs": 1842
  }
}

Analyze Email (Legacy)

Backward-compatible endpoint for email-only analysis. Use /api/analyze for new integrations.

POST /api/analyze-email

{ "email": "Customer email body text..." }

Batch Analyze

Analyze up to 50 interactions in a single request. Results are returned sorted by risk score (highest first).

POST /api/batch-analyze

{
  "emails": [
    { "id": "email-001", "content": "First email...", "from": "alice@co.com" },
    { "id": "email-002", "content": "Second email...", "from": "bob@co.com" }
  ]
}

Response includes a results array sorted by risk score, plus summary with counts by level and the highest risk item.

Inbound Email Webhook

Webhook endpoint for email forwarding services (SendGrid Inbound Parse, Cloudflare Email Workers, etc.).

POST /api/inbound-email

// JSON format
{ "text": "Email body...", "subject": "RE: Issue", "from": "customer@co.com" }

// Also accepts SendGrid multipart form data

Risk Levels

Level	Score		Description
critical	80-100	●	Immediate intervention required. Legal threats, formal complaints.
high	60-79	●	Respond today. Competitor evaluation, escalation demands.
medium	30-59	●	Address this week. Growing frustration, repeat contacts.
low	0-29	●	Routine. No significant risk signals detected.

Signal Types

RiskDetect tracks 30+ risk signals across these categories:

Legal: Attorney mention, BBB complaint, regulatory threat, contract review
Competition: Competitor named, comparison language, switching intent
Escalation: Manager request, deadline ultimatum, repeated contact
Sentiment: Declining tone, frustration language, all-caps, profanity
Behavioral: Usage decline, team downsizing, contract non-renewal
Financial: Billing dispute, refund request, overcharge claim

Rate Limits

Plan	Limit	Notes
Starter	100/mo	~3/day
Growth	1,000/mo	~33/day, burst up to 50/hr
Scale	Unlimited	Fair use policy, 100/min burst

Rate limit headers included in every response: X-RateLimit-Remaining, X-RateLimit-Reset.