NOCAP™ · Media Verification

Truth online is no longer obvious.

NOCAP analyzes suspicious images, video, audio, and screenshots for manipulation signals—then shows you the evidence behind the result.

Signals, not proof. Evidence, not certainty.

NOCAP · SCAN RESULT
REAL 95% confidence · High — not proof of authenticity
Metadata consistencyBasic EXIF fields are present.
Model AI scoreRaw model score: 0.001842
Structural analysisNon-round JPEG dimensions are consistent with a camera capture.
Evidence-firstEvery verdict ships with the signals behind it.
Capture device fieldsMake / Model present — consistent with a phone camera.
Honest caveats“Software metadata suggests the file may have been exported or edited.”

The problem

A confident answer is not always an honest answer.

Synthetic media is improving

Generated images, video, and voices are increasingly difficult to distinguish from real captures — even for careful eyes.

Compression removes evidence

Every re-upload, screenshot, and re-encode strips away the forensic traces a detector relies on.

Context gets lost

Media travels without its history. By the time something reaches you, where it came from — and what happened to it — is gone.

Product

A responsible second opinion for suspicious media.

NOCAP is a media-verification assistant built to give people a responsible second opinion when something online does not feel right. It examines the relevant signals and communicates what it observed — it does not claim to prove authenticity.

Image analysis

Examines metadata, compression, texture, and model-based signals in photos and generated images.

Video analysis

Samples frames and evaluates per-frame signals together with temporal and compression context.

Audio analysis

Reviews recorded or uploaded audio for spectral and consistency signals associated with synthetic speech.

Screenshot & message analysis

Reads visible conversation text and flags common scam-pressure and manipulation patterns.

The honest part

Why detection is difficult

The expected experience

  • Upload media — drop a file in a box.
  • Receive a binary answer — “real” or “fake,” instantly.
  • Treat confidence as certainty — 98% sounds like proof.

The real challenge

  • Recompression can remove useful evidence.
  • Edited media may contain both authentic and synthetic elements.
  • Different creation tools can leave different signals.
  • Weak evidence should not support a strong accusation.

NOCAP is designed to communicate when the evidence is insufficient.

Verdict philosophy

A detector should be allowed to say “I don’t know.”

Every NOCAP scan resolves to one of five verdicts. Each one is a statement about evidence — never a claim of proof.

REAL

Available signals lean toward authentic media. This is not proof of authenticity.

MIXED

Conflicting, blended, or inconsistent signals were observed.

CAP DETECTED

Manipulation or synthetic-media signals were detected. This is not proof of fabrication.

UNCERTAIN

The available evidence is too weak or conflicting for a stronger conclusion.

UNKNOWN

The media could not be meaningfully assessed.

Abstention helps prevent weak evidence from being presented as certainty.

Interactive sample walkthrough

See a real saved result.

Interactive sample walkthrough These examples were previously analyzed using NOCAP. The results below are saved demonstrations from an internal evaluation build (July 2026). No media is uploaded or scanned on this website.
How it works

Four stages. No black box.

  1. Choose media

    Record or select supported content inside the mobile application.

  2. Analyze signals

    NOCAP evaluates relevant forensic, contextual, and consistency signals.

  3. Review the result

    See the verdict, confidence, and evidence observations.

  4. Make your own judgment

    Use NOCAP as one source of information—not the final authority.

Product experience
Privacy & trust

Built to inform judgment, not replace it.

No public upload portal

This website does not accept arbitrary media for analysis.

Evidence over accusations

NOCAP explains signals and limitations.

Uncertainty is communicated

Weak evidence is not disguised as certainty.

Responsible use

A NOCAP result should never be the sole basis for a high-stakes decision or public accusation.

Vision

A verification layer for the synthetic internet.

NOCAP’s long-term vision is to make responsible media verification accessible wherever suspicious content appears—without pretending that a probabilistic system can replace human judgment, provenance, or professional investigation.

Mobile today Broader beta Future visionTrusted web workflows Future visionVerification integrations
Roadmap

Where NOCAP is headed

Now

  • Closed iOS alpha
  • Image, video, audio, and screenshot workflows
  • Evidence-first result experience
  • Structured product feedback

Future

  • Responsible web-based verification workflows
  • Deeper provenance and contextual analysis
  • Tools for professional and organizational use

Roadmap subject to change.

Origin

Why NOCAP exists

“When evidence is weak, the product should say so.”

NOCAP began with a practical frustration: as generated media became more convincing, many detection experiences still reduced complicated evidence to an overconfident yes-or-no answer. Built by Dante Beltran through Anomaly X Innovations, NOCAP combines experience across systems, cloud infrastructure, identity, security operations, and product development with that simple principle.

See what responsible media verification can look like.