AI authenticity intelligence, in your browser

Check AI images and videos.Review the evidence.

DetectFake screens photos and sampled video frames for signs of AI generation and manipulation. Review the signals alongside the original source: a detector result is an estimate, not proof that media is real or fake.

3 free daily scans No paid AI per scan Explainable results

Authenticity report

Illustrative report

Local scan

Example verdict

AI Generated

85%

confidence

Fast

Local inference

Explainable

Signal evidence

Private

No paid gateway

Authenticity scan

Analyze an image or video

Local-first Signal-based screening

Why choose DetectFake for authenticity analysis

Private by design

Normal scans run locally in the browser or app without a paid per-scan AI gateway.

Evidence, not just a label

See confidence, provenance and forensic signals behind every result.

Fast local workflow

Upload, analyze and review suspicious media without sending every scan to an external AI service.

Explainable detection

AI image and video detection with clear evidence

DetectFake checks photos, screenshots, artwork and sampled video frames for patterns linked to generative AI, deepfakes and manipulation. Each scan returns a probabilistic verdict, confidence score and the signals that influenced it.

Visual forensics

Texture, noise, compression and object-boundary cues.

Video checks

Frame sampling and consistency signals across a clip.

Portrait review

Face-region evidence such as skin texture and structural cues.

Probabilistic verdicts

Clear confidence rather than pretending every file is certain.

How to check suspicious media

  1. Choose the original image or video when available. Screenshots and messaging-app compression can remove useful detail.
  2. Run the scan and read the supporting signals, not just the verdict. Video analysis samples frames; it does not verify every frame or the audio.
  3. Check where the file first appeared, compare earlier versions, and ask the source for the original before making an accusation or sharing it as fact.

Start with the AI image verification guide or the viral video verification guide for a step-by-step review.

Record your sources, evidence and next steps in the free verification worksheet, then download your notes as a text report. Notes stay in your browser until you download them; they are not saved when you leave the page.

What a result can and cannot tell you

A camera photo can trigger AI-like signals after editing, scanning or compression. An AI-generated image can also escape detection. A confidence score is not a guarantee of accuracy for your file.

A processed document or WhatsApp image describes how a file may have been handled, not proof of its origin. Do not use a scan alone to establish someone's identity, authenticate a document or resolve a dispute.

Read AI detector accuracy and limitations and our analysis methodology before relying on a verdict.

DetectFake provides probabilistic forensic assessments and is not a substitute for expert review. Results should be interpreted alongside provenance data (C2PA, EXIF, chain of custody) before high-stakes decisions.