Aethyl: AI Video Detector

Verify reality. Trust nothing. Paste any TikTok, Instagram, X, YouTube, or Reddit link and get a Trust Score from three AI detectors voting together.

Aethyl tells you whether a video was made by a human or AI.

Paste a link from TikTok, Instagram, X, YouTube, or Reddit and get a Trust Score in seconds. Under the hood, multiple AI detection models vote together to produce a verdict — more accurate than any single detector alone.

WHAT YOU GET • Trust Score from 0 (likely AI) to 100 (likely real) • Per-frame timeline showing where the video is suspicious • Model-by-model verdict breakdown — see what each detector thought • Community discussion on every scan • Share verdicts with one tap

WHY IT'S DIFFERENT • Multiple detectors. Most apps use one model. We use three (Gemini, GPT-5 Mini, and Grok) and weight their votes for better accuracy. • Works on real social content. Paste links from the platforms you actually use — no downloads, no file conversions, no manual frame extraction. • No tracking. We don't sell data, we don't build ad profiles, we don't participate in cross-app tracking networks. • Free tier: three scans per day, videos up to 30 seconds. Pro tier: one hundred scans per week, videos up to 5 minutes.

HOW IT WORKS

1. Paste a social media link
2. Aethyl downloads the media, extracts frames, and sends them to three AI providers in parallel
3. Each provider returns per-frame probabilities
4. Votes are weighted and aggregated into a Trust Score
5. You can comment, vote on accuracy, and share

SUPPORTED PLATFORMS • TikTok • Instagram (posts and reels) • X / Twitter • YouTube (shorts and long-form) • Reddit (video posts and image galleries)

PRIVACY We collect only what we need to run the app: your email for sign-in, your handle, your scan history, and your comments. Media you submit is analyzed and stored so future users can benefit from cached results. You can delete your account anytime from Settings.

ACCURACY DISCLAIMER AI detection is imperfect. Results are probabilistic estimates, not proof. Results may be wrong in either direction — don't rely on Aethyl as the sole basis for legal, journalistic, or safety-critical decisions.

Questions? https://aethyl.ai/support Privacy Policy: https://aethyl.ai/privacy Terms of Service: https://aethyl.ai/terms