Media Provenance

Trace the origin and authenticity of images and videos. Detects deepfakes, verifies C2PA digital watermarks, checks metadata consistency, and cross-references against known media databases.

Endpoint

POST https://signal-stack-ten.vercel.app/v1/verify/media

Request parameters

Parameter	Type	Required	Description
`url`	string	Yes*	URL of the image or video to analyze
`file`	string	Yes*	Base64-encoded file data (alternative to URL)
`check_deepfakes`	boolean	No	Run deepfake detection model (default: `true`)
`check_provenance`	boolean	No	Trace image provenance across the web (default: `true`)
`check_watermarks`	boolean	No	Extract and verify C2PA watermarks (default: `true`)

* Either url or file is required.

Response schema

Field	Type	Description
`id`	string	Unique request identifier
`trust_score`	number	Overall authenticity confidence (0.0–1.0)
`authenticity`	string	`authentic`, `likely_authentic`, `suspicious`, `manipulated`
`deepfake_score`	number	Probability of deepfake manipulation (0.0–1.0)
`provenance`	array	Timeline of known appearances of this media
`provenance_source`	string	Earliest known source of the media
`watermark`	object	C2PA or other digital watermark data
`metadata`	object	Extracted EXIF and other metadata
`evidence`	array	Source evidence chain
`latency_ms`	number	Processing time in milliseconds

Request example

{
  "url": "https://example.com/photo.jpg",
  "check_deepfakes": true,
  "check_provenance": true,
  "check_watermarks": true
}

Response example

{
  "id": "vrf_med_c9d0e1f2",
  "trust_score": 0.12,
  "authenticity": "manipulated",
  "deepfake_score": 0.94,
  "provenance": [
    {
      "url": "https://original-source.com/photo.jpg",
      "first_seen": "2024-06-15T08:30:00Z",
      "source": "news_archive"
    },
    {
      "url": "https://example.com/photo.jpg",
      "first_seen": "2025-01-20T14:22:00Z",
      "source": "web_crawl"
    }
  ],
  "provenance_source": "https://original-source.com/photo.jpg",
  "watermark": null,
  "metadata": {
    "width": 1920,
    "height": 1080,
    "format": "JPEG",
    "exif": {
      "make": null,
      "model": null,
      "datetime_original": null,
      "gps": null
    }
  },
  "evidence": [
    {
      "source": "deepfake_detection",
      "finding": "GAN-based manipulation artifacts detected in facial region (p=0.94)",
      "authority_score": 0.93
    },
    {
      "source": "provenance_analysis",
      "finding": "Image first appeared June 2024; EXIF metadata has been stripped",
      "authority_score": 0.87
    }
  ],
  "dimensions": {
    "source_authority": 0.90,
    "cross_reference": 0.76,
    "recency": 0.65,
    "llm_analysis": 0.89
  },
  "latency_ms": 2100,
  "credits_used": 4
}

Code examples

Python

import signalstack

client = signalstack.SignalStack(api_key="ssk_live_...")

result = client.verify.media(
    url="https://example.com/photo.jpg",
    check_deepfakes=True,
)

print(f"Authenticity: {result.authenticity}")
print(f"Deepfake score: {result.deepfake_score}")
print(f"Provenance: {result.provenance_source}")

cURL

curl -X POST https://signal-stack-ten.vercel.app/v1/verify/media   -H "Authorization: Bearer $SIGNALSTACK_API_KEY"   -H "Content-Type: application/json"   -d '{
    "url": "https://example.com/photo.jpg",
    "check_deepfakes": true
  }'

TypeScript

import { SignalStack } from "@signalstack/sdk"

const client = new SignalStack({ apiKey: process.env.SIGNALSTACK_API_KEY! })

const result = await client.verify.media({
  url: "https://example.com/photo.jpg",
  checkDeepfakes: true,
})

console.log(`Authenticity: ${result.authenticity}`)
console.log(`Deepfake score: ${result.deepfakeScore}`)

Next steps

Explore Claim Verification
Learn about Trust Scoring