Is it real?
Get the analysis.
Six forensic detectors and four neural specialists scan every pixel, frequency, and byte of metadata — pooling their analysis into a single, unified trust score. Your media is processed securely in memory and instantly destroyed, ensuring total privacy.
- forensic detectors
- 6
- forensic detectors
- neural specialists
- 4
- neural specialists
- pixels retained
- 0
- pixels retained
- analyses per scan
- 10
- analyses per scan
Native iOS & Android Apps
Zero-log media scanning directly from your device.
Full-Frame Video Analysis
Temporal deepfake detection and frame-by-frame anomaly scoring.
Analysis report
Synthetic
confidence 96.9%
Five trust levels. Zero guesswork.
Every score is a pooled probability across all detectors — never one model's opinion.
signed C2PA credentials
0–30% AI probability
30–55% AI probability
55–80% AI probability
80–100% AI probability
How the pipeline works
Like a multi-layered security grid — no single anomaly triggers an alert, but combined cryptographic and visual data provide absolute clarity.
Zero-Log Fingerprinting
SHA-256 and perceptual hashing run locally before a single byte leaves your device. If the network has already analyzed the media, it returns an instant score — no upload required.
Ensemble Detection
Metadata forensics, spectral falloff, compression analysis, and sensor-noise mapping — plus four neural specialists scanning across independent visual domains to catch what the human eye misses.
Weighted Consensus
A sophisticated log-odds pooling architecture. Weak signals are filtered out, while compounding evidence from highly confident, uncorrelated models is fused into a single authenticity score.
The next image you doubt — scan it.
Free while in development. Sign in with Google and get a full forensic report in seconds.
Start scanning