Real Photos,
Certified by Physics.
The world's first Optical Attestation Engine. Instead of chasing ever-mutating generative AI artifacts, we mathematically verify the immutable physical laws of camera optics, sensor silicon, and corneal reflection.
⚡ 100% Free · 🔒 Zero Photos Saved · 📱 Early Testing on Google Play.
Building a platform or dating app? Explore Enterprise REST API & 30-Day Pilot →
Interactive Demo: Real vs. AI
Taken with a Real Camera
Eye reflections, light angles, and camera grain match real-world physics.
Professional forensics.
Consumer simplicity.
The Shazam of Physical Reality
Standard detectors rely on pixel pattern matching that degrades whenever a new AI generator launches. EktaCore is the Shazam of authenticity: we verify 37 optical physical properties, including corneal reflection concordance and sensor shot noise, to mathematically attest genuine physical camera provenance.
Attestation & Physical Veto
Built on TRIZ Principle 13 (Inversion), EktaCore pairs positive camera manifold attestation ($D_M \le \tau$) with shifted boundary tripwires. If a capture violates corneal symmetry, Fourier phase consistency, or sensor shot noise, it is denied authenticity in sub-second CPU time.
No Cryptic Guesses
No confusing "58% fake" black boxes. EktaCore generates clear, auditable physical certificates explaining exactly how the photo proved its camera provenance.
Visual Heatmaps
Inspect exact anomaly locations across facial features, showing precisely where synthetic lighting and reflections break down.
Zero-Trace Privacy
Images are processed strictly in volatile memory and immediately wiped. Zero disk storage, zero training retention. EEA & GDPR compliant.
Ultra-Low Latency
Sub-second forensic verification designed for mobile camera captures and stateless high-throughput REST API integration.
Three taps. Real answers.
Upload any face photo from your gallery, camera, or a screenshot from social media.
Select a photo
Choose any photo with a face from your gallery, camera, or a screenshot from social media or messaging apps.
Forensic analysis
Your image is encrypted in transit and analyzed in memory on EU servers. We evaluate optical properties, such as eye reflections and shadow consistency across the face.
Certified Attestation
Get an affirmative Physical Camera Certificate. If verified, enjoy green-shield certainty; if physics laws are broken, view the exact optical breakdown.
Physics doesn't lie.
AI generators evolve every month. The laws of camera optics don't.
| Typical AI Detectors | EktaCore | |
|---|---|---|
| Approach | Chasing synthetic artifacts | The Shazam of Authenticity (Camera Manifold) |
| Explainability | Cryptic black box (e.g. 58%) | Affirmative physical attestation |
| New AI models | Breaks on new releases | Camera sensor physics never changes |
| False Alarm Rate | 10-25% (unpredictable) | 3.6% (96.4% True Verification Rate) |
| Cross-Model Testing | Single-generator bias | 80+ generative AI engines tested |
| Privacy | Images stored | Zero retention (In-RAM execution) |
| Speed | 10-30 seconds | < 0.2s Manifold Verification |
| Infrastructure | Heavy GPU clusters | Lightweight CPU, scales instantly |
| EU AI Act | Not addressed | Article 50 ready |
Built for trust.
Online Dating
Verify profile photos before meeting
Journalism
Verify source images before publishing
HR & Recruiting
Screen candidate profile photos
Parents
Check suspicious social media profiles
Security
Due diligence and KYC verification
Research
Study synthetic media at scale
The "Shazam" of Authenticity.
Built for Enterprise.
Integrate our stateless, physics-based detection engine directly into your platform with sub-second REST API latency.
EU AI Act (Article 50) requires platforms to detect and label synthetic media. EktaCore is built for compliance: GDPR-compliant by design, EEA-hosted, with zero data retention.
Instant sandbox credentials · Sub-second REST latency · EU AI Act Art. 50 ready
Frequently Asked Questions
How does EktaCore detect deepfakes?
EktaCore measures 37 optical properties in a portrait, including eye reflections, shadow geometry, and light source direction. Because cameras follow physical laws, optical verification stays reliable even as new AI models are released.
Is EktaCore free?
Yes. EktaCore offers 5 free scans per day on Android. You can earn additional scans by helping calibrate the forensic engine. No account or credit card required.
Does EktaCore store my photos?
No. Images are processed in memory on EU servers and immediately discarded. Zero photos are stored. Only anonymous optical metrics help calibrate detection thresholds. EktaCore is GDPR compliant by design.
Can EktaCore detect photos from ChatGPT, Midjourney, or DALL-E?
Yes. Generative models struggle to produce consistent 3D lighting, especially matching eye reflections and natural shadow distribution across facial features. EktaCore detects these physical inconsistencies.
Does EktaCore scan photos of pets, animals, or objects?
No. EktaCore is specialized for human facial forensics and optical light path verification across human features. If an image contains an animal (such as a dog or cat) or a non-human object, the system automatically detects the non-human subject and prompts you to upload a photo containing a human face.
Is EktaCore available for enterprise or API use?
Yes. EktaCore provides a stateless REST API for B2B integration with sub-second latency, EU AI Act Article 50 compliance, and zero data retention. Request a 30-day enterprise API pilot above →
Can EktaCore analysis be used as legal proof or in court?
No. EktaCore provides probabilistic indicators for preliminary screening and research. Results are not judicial evidence, expert witness testimony, or definitive proof of authenticity. Always seek certified forensic experts for formal legal or investigative proceedings.
Stop Guessing.
Start Verifying.
EktaCore is live in Early Testing on Google Play. Apply with your Google Play email to receive an invitation to test the app for free.
By applying, you agree to our Privacy Policy and Terms of Service. Verified Google Play account required.