Sys.Directive // 0x4A/Editorial Forensics

Journalism & Source Verification

Protect your newsroom's credibility. Sightova scores how likely an image is to be AI-generated, checks for invisible SynthID and OpenAI watermarks, and shows what its C2PA credentials and metadata say, so your editors can publish with evidence.

REST API + API KEY
C2PA CREDENTIALS CHECKED
IMAGES DELETED AFTER ANALYSIS

Verification Capabilities

QUERY: SELECT * FROM editorial_tools
JRN-META-01
FORENSIC MODULE

EXIF Metadata Reading

Read EXIF metadata, including capture time, camera model, and GPS coordinates when the file carries them, so a journalist can compare them with what the source claims. Even when metadata has been stripped, the model still reads the pixels.

EXIF EXTRACTIONCAPTURE TIMECAMERA MODELGPS (IF PRESENT)
JRN-C2PA-02
FORENSIC MODULE

C2PA Content Credentials

Validate Content Credentials when an image carries them. A valid credential can show which camera or tool made the image and what edits were recorded, so you can point to the tool with evidence.

CONTENT CREDENTIALSSIGNER INFORECORDED EDITS
JRN-DEEP-03
FORENSIC MODULE

AI-Generated Image Detection

Run pixel-level analysis to score how likely an image is to be AI-generated, and see the closest generator style. When a watermark or C2PA credentials are present, they name the exact tool.

AI PROBABILITYGENERATOR HINTHEATMAP
JRN-DATA-04
FORENSIC MODULE

Hosted in Germany, Deleted After Analysis

Submitted images are processed on our servers in Germany, deleted after analysis, and not used for training.

HOSTED IN GERMANYDELETED AFTER ANALYSISNO TRAINING
JRN-API-05
FORENSIC MODULE

REST API for Editorial Tools

Call the REST API from your own editorial tools or scripts to flag likely AI-generated images before they reach the desk, with results in seconds.

REST APIAPI KEYRESULTS IN SECONDS
JRN-EVD-06
FORENSIC MODULE

Evidence for Editorial Review

Each result gives editors a score, a verdict, any C2PA or watermark findings, and, on Starter and higher plans, a heatmap of the regions that look AI-edited, to use in your own verification notes. You set the threshold, and the editorial call stays with your desk.

SCORESHEATMAPYOUR THRESHOLD
// META-ANALYSIS

Evidence for Fact-Checking

In the era of rapid disinformation, human intuition is not always enough. Sightova combines a read of C2PA content credentials and invisible watermarks with pixel-level AI detection, giving editors evidence to weigh, even when metadata has been stripped. It is tested on 250k+ benchmark images and 50+ generative models, with results published in the testing methodology.

  • C2PA manifest validation and signer details, when credentials are present
  • EXIF metadata reading, including GPS and capture time when present
  • Invisible watermark checks (Google SynthID, OpenAI)
RESPONSE /api/v1/detect/ai
{
  "request_id": "req_5d2e8a41b7c9",
  "detection_type": "image moderation",
  "result": {
    "ai_probability": 0.99,
    "real_probability": 0.01,
    "confidence": 0.97,
    "verdict": "ai-generated"
  }
}

Check Your Source Images

Give your editorial desk a second opinion on images in seconds: an AI probability score, watermark, C2PA and metadata findings, and a heatmap to guide a closer look.