Understand the score
The percentage is detector confidence, not a guarantee. Never use it alone to accuse a person or make a legal, safety or disciplinary decision.
ShieldHer AI Deepfake Detection Centre
Screen a photo or video for signs of synthetic generation.
ShieldHer Signal Fusion combines Hive V3 media analysis with an independent on device cross-check. Results are screening signals, not proof that media is real or fake.
Three signal checks when availablePhoto, video + audio screeningConflict safeguard holds mixed results
Choose media to examine
Use the clearest original file available. Avoid screenshots when the original image or video is available.
JPG, PNG, WebP up to 20 MB · MP4, MOV, WebM up to 40 MBChoose a photo or video to begin.
Choose your privacy level
The lab scan sends the selected file to Hive only after consent, then performs a separate local cross-check. ShieldHer AI does not save the file. The private quick scan remains in your browser.
Verify twice
Check the original source and context, inspect provenance, and use a second independent verification method when the result matters.
ShieldHer model transparency: ShieldHer Signal Fusion 1.2 is our original decision layer. It compares Hive V3 with an independent local synthetic image detector and deliberately marks strong disagreement as inconclusive. Hive keeps a separate face manipulation and audio check. ShieldHer has not yet completed independent benchmarking, so we do not claim it outperforms every detector.
Read the model documentation and limits ↗