Decision layer · Built on itsreal.media signals

Risk Score API

Turn authenticity signals into one operational number. Automated allow / review / block decisions at scale.

Send a claim text and photos or videos and get back a single 0–100 risk score. The Risk Score API runs Frame:Detect, Content Credentials verification, and claim-image consistency analysis automatically, so your rules engine can act immediately.

Request access View API docs →
0–100 Normalized risk scale
< 200ms p95 response time
3 Upstream signal sources
Risk Score API · Live result Live
0
LOW 0–20 REVIEW 21–60 HIGH 61–100
Frame:Origin 0.94
Frame:Edit manipulation detected
Content Credentials missing
risk_band: high action: escalate
How it works

What the score is

The Risk Score API is…

Send only the claim text and photos or videos — nothing else required
Runs Frame:Detect, Content Credentials, and claim-image consistency analysis automatically
One consistent 0–100 number per submission
Configurable thresholds — you define what LOW, MEDIUM, and HIGH mean for your business
Fully auditable — every score includes the rationale and contributing factors
Layer 01 · Authenticity
Frame:Origin likelihood
Frame:Edit likelihood
Content Credentials status
Layer 02 · Client input
Photos or videos
Claim text
Output · Fused score
Normalized 0–100
Risk band
Rationale
Recommended action
Input to output

Live API request

Request · cURL 
curl -X POST https://api.itsreal.media/v1/risk-score \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "claim_text": "Water damage in kitchen, pipe burst under sink...",
    "images": [
      "https://upload.example.com/img_8f3kx92.jpg",
      "https://upload.example.com/img_8f3kx93.jpg"
    ]
  }'
Response · JSON 
{
  "id": "rsk_01hx9k2m7q",
  "score": 72, // 0 = authentic, 100 = synthetic
  "risk_band": "high",
  "action": "escalate",
  "rationale": {
    "contributing_factors": [
      "ai_generated_image_detected",
      "content_credentials_absent",
      "claim_text_image_inconsistency"
    ]
  },
  "signals_used": [
    "frame_origin",
    "frame_edit",
    "content_credentials",
    "claim_text_analysis"
  ],
  "processed_at": "2025-11-04T14:22:31Z"
}
Claim text + images Input
→
Forensic analysis Frame:Detect
→
Risk score 72%
→
Band HIGH
Thresholds

Your thresholds. Your rules.

Every business defines risk differently. Configure the score bands and actions that match your operational playbook.

LOW
0 – 20
Auto-approve
Insurance: fast-track claim
E-commerce: instant listing approval
MEDIUM
21 – 60
Manual review
Insurance: request supporting docs
E-commerce: seller verification
HIGH
61 – 100
Block / escalate
Insurance: SIU referral
E-commerce: suspension review
Context-aware

Just send the claim and the images.

No complex integrations. Send the claim text and photos or videos. we handle all forensic analysis and return a single risk score.

Base layer — Authenticity

Frame:Origin AI-generation likelihood, Frame:Edit manipulation likelihood, and Content Credentials verification status. These signals form the raw authenticity score.

Client input — What you send

You only need to send the claim text and the associated photos or videos. We handle all forensic analysis, signal extraction, and scoring automatically.

images
The photos submitted with the claim — or a single PDF, whose embedded photos are extracted (up to 5). Supported formats: JPEG, PNG, WebP, AVIF, PDF.
claim_text
The text description of the claim. Used for cross-referencing consistency between the written narrative and the visual evidence.

Who uses Risk Score API

One API, many operational contexts.

Insurance / claims
Fraud detection at FNOL
E-commerce / marketplaces
Listing verification
Trust and safety
Content moderation
Compliance teams
Regulatory audit trail

Check your dataset for integrity.

Send us a sample of your claims data. images, videos, and claim text, and we will return scored results with a full integrity report and threshold recommendation.

Request access View API docs →